I N T R O D U C T I O N T O M O D E R N C L I M A T E
C H A N G E , S E C O N D E D I T I O N
This is an invaluable textbook for any introductory survey course on the
science and policy of climate change, for both non–science majors and
introductory science students. The second edition has been thoroughly
updated to reflect the most recent science from the latest IPCC reports, and
many illustrations include new data. The new edition also reflects
advances in the political debate over climate change. Unique among
textbooks on climate change, this text combines an introduction to the
science with an introduction to economic and policy issues, and it focuses
closely on anthropogenic climate change. It contains the necessary
quantitative depth for students to properly understand the science of
climate change. It supports students in using algebra to understand simple
equations and to solve end-of-chapter problems. Supplementary online
resources include a complete set of PowerPoint figures for instructors,
solutions to exercises, videos of the author’s lectures, and additional
computer exercises.
Andrew Dessler is a climate scientist who studies both the science and
politics of climate change. His scientific research revolves around climate
feedbacks, in particular how water vapor and clouds act to amplify
warming from the carbon dioxide that human activities emit. During the
last year of the Clinton administration, he served as a senior policy analyst
in the White House Office of Science and Technology Policy. Based on
his research and policy experience, he has authored two books on climate
change: this textbook and The Science and Politics of Global Climate
Change: A Guide to the Debate (co-written with Edward Parson; second
edition published in 2010). This textbook won the 2014 American
Meteorological Society Louis J. Battan Author’s Award. In recognition of
2

his work on outreach, in 2011 he was named a Google Science
Communication Fellow. He is presently a professor of atmospheric
sciences at Texas A&M University. His educational background includes a
B.A. in physics from Rice University and a Ph.D. in chemistry from
Harvard University. He also undertook postdoctoral work at NASA’s
Goddard Space Flight Center and spent nine years on the research faculty
of the University of Maryland.
3

INTRODUCTION TO
MODERN CLIMATE
CHANGE
Second Edition
Andrew Dessler
Texas A&M University
4

University Printing House, Cambridge CB2 8BS, United Kingdom
One Liberty Plaza, 20th Floor, New York, NY 10006, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia
4843/24, 2nd Floor, Ansari Road, Daryaganj, Delhi – 110002, India
79 Anson Road, #06-04/06, Singapore 079906
Cambridge University Press is part of the University of Cambridge.
It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.
www.cambridge.org
Information on this title: www.cambridge.org/9781107480674
© Andrew Dessler 2016
This publication is in copyright. Subject to statutory exception and to the provisions of relevant
collective licensing agreements, no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2016
5th printing 2017
Printed in the United Kingdom by TJ International Ltd. Padstow, Cornwall
A catalog record for this publication is available from the British Library.
Library of Congress Cataloging in Publication Data
Dessler, Andrew Emory.
Introduction to modern climate change / Andrew Dessler, Texas A&M University. –
[Second edition].
pages cm
Includes bibliographical references and index.
ISBN 978-1-107-09682-0
5

http://www.cambridge.org

http://www.cambridge.org/9781107480674

1. Climatic changes. 2. Climatic changes – Government policy. I. Title.
QC903.D46 2016
551.6–dc23 2015014701
ISBN 978-1-107-09682-0 Hardback
ISBN 978-1-107-48067-4 Paperback
Additional resources for this publication at www.andrewdessler.com
Cambridge University Press has no responsibility for the persistence or accuracy of URLs for
external or third-party Internet Web sites referred to in this publication and does not guarantee
that any content on such Web sites is, or will remain, accurate or appropriate.
6

http://www.andrewdessler.com

For Michael and Alex
7

Contents
Preface
Acknowledgments
1 An introduction to the climate problem
2 Is the climate changing?
3 Radiation and energy balance
4 A simple climate model
5 The carbon cycle
6 Forcing, feedbacks, and climate sensitivity
7 Why is the climate changing?
8 Predictions of future climate change
9 Impacts of climate change
10 Exponential growth
11 Fundamentals of climate change policy
12 Mitigation policies
13 A brief history of climate science and politics
14 Putting it together: A long-term policy to address climate
change
8

References
Index
9

Preface
Future generations may well view climate change as the defining issue of
our time. The worst-case scenarios of climate change are truly terrible, but
even middle-of-the-road scenarios portend environmental change without
precedent for human society. When future generations look back on our
time in charge of the planet, they will either cheer our foresight in dealing
with this issue or curse our lack of it.
Yet despite the stakes, the world has done basically nothing to
address this risk. The reasons are obvious: The threat of climate change is
really a threat to future generations, not the present one, so actions taken
by our generation will mostly benefit them and not us. Moreover, such
actions may be expensive – reducing emissions means rebuilding our
energy infrastructure, and we have no idea how much that will cost. In
such a situation, it is easiest to do nothing and wait for disaster to strike –
which is why dams are frequently built after the flood, not before.
Nevertheless, pushing this problem off onto future generations is a poor
strategy. The impacts of climate change are global and mainly irreversible;
by the time we have unambiguous evidence that the climate is changing
and its impacts are serious, it will be too late to avoid these serious
impacts. The only hope that future generations have to avoid serious
climate change is us.
I fully believe that the cornerstone of good policy is an electorate that
is educated on the issues, and this belief provided me the motivation for
writing this book. The goal of this book is to cover the human-induced
10

climate change problem from stem to stern, covering not just the physics
of climate change but also the economic, policy, and moral dimensions of
the problem. This sets it apart from most other climate change books,
which typically do not have a tight focus on human-induced climate
change or do not cover the nonscience aspects of the problem.
Such complete coverage of the climate change problem is essential.
The science clearly underlies all discussion of the problem, and an
understanding of the science is essential to an understanding of why so
many people are so worried about it. Climate change, however, is no
longer just a scientific problem. Virtually every government in the world
now accepts the reality of climate change, and the debate has, to a great
extent, moved on to policy questions, including the economic and ethical
issues. Thus, one must also understand nonscience aspects of the problem
to be truly informed on this issue.
The first seven chapters of the book focus on the science of climate
change. Chapter 1 defines the problem and provides definitions of weather,
climate, and climate change. It also addresses an issue that most textbooks
do not have to address: why the reader should believe this book as opposed
to Web sites and other sources that give a completely different view of the
climate problem. Chapter 2 explains the evidence that the Earth is
warming. The evidence is so overwhelming that there is little argument
anymore over this point, and my goal is for readers to come away from the
chapter understanding this.
Chapter 3 covers the basic physics of electromagnetic radiation
necessary to understand the climate. I use familiar examples in this
chapter, such as glowing metal in a blacksmith shop and the incandescent
light bulb, to help the reader understand these important concepts. In
Chapter 4, a simple energy-balance climate model is derived. It is shown
how this simple model successfully explains the Earth’s climate as well as
11

the climates of Mercury, Venus, and Mars. Chapter 5 covers the carbon
cycle, and feedbacks, radiative forcing, and climate sensitivity are all
discussed in Chapter 6. Finally, Chapter 7 explains why scientists are so
confident that humans are to blame for the recent warming that the Earth
has experienced.
Chapter 8 begins an inexorable shift from physics to nonscience
issues. It discusses emissions scenarios and the social factors that control
them, as well as what these scenarios mean for our climate over the next
century. Chapter 9 covers the impacts of these changes on humans and on
the world in which we live. Chapter 10 covers exponential math.
Exponential growth is a key factor in almost all fields of science, as well
as in real life. In this chapter, I cover the math of exponential growth and
explain the concept of exponential discounting. I also touch briefly on the
social cost of carbon.
Starting with Chapter 11, the discussion is entirely on the policy
aspects of the problem. Chapter 11 discusses the three classes of responses
to climate change, namely adaptation, mitigation, and geoengineering, and
their advantages, disadvantages, and trade-offs. The most contentious
arguments over climate change policy are over mitigation, and Chapter 12
discusses in detail the two main policies advanced to reduce emissions:
carbon taxes and cap-and-trade systems.
Chapter 13 provides a brief history of climate science and a history of
the political debate over this issue, including discussions of the United
Nations’ Framework Convention on Climate Change and the Kyoto
Protocol. Finally, Chapter 14 pulls the last three chapters together by
discussing how to decide which of our options we should adopt,
particularly given the pervasive uncertainty in the problem.
Overall, it should be possible to cover each chapter in three hours of
lecture. This makes it feasible to cover the entire book in one fifteen-week
12

semester. At Texas A&M, the material in this book is being used in a one-
semester class for nonscience majors that satisfies the university’s science
distribution requirement. Thus, it is appropriate for undergraduates with
any academic background and at any point in their college career.
Any serious understanding of climate change must be quantitative.
Therefore, the book assumes a knowledge of simple algebra. No higher
math is required. The book also assumes no prior knowledge of any field
of science, just an open mind and a willingness to learn. To aid in the
student’s development of a numerate understanding of the climate, there
are quantitative questions at the end of many of the chapters, and every
chapter also has more open-ended, qualitative questions. In addition, there
is a chapter summary at the end of each chapter that reviews and
summarizes the most important takeaway messages from the chapter. A
list of important terms is also provided at the end of each chapter. I’ve put
additional readings, video recordings of my lectures, and computer
exercises on my Web site, www.andrewdessler.com.
This is not an advocacy book. This is not to say that I do not have
opinions. I do, and strong ones. I recognize, though, that shrill advocacy is
frequently less effective than a dispassionate presentation of the facts.
Thus, my strategy in this book is to simply explain the science and then lay
out the possible solutions and trade-offs among them. I firmly believe that
an unbiased assessment of the facts will bring the majority of people to see
things the way I do: that climate change poses a serious risk and that we
should therefore be heading off that risk by reducing our emissions of
greenhouse gases.
Every year that our society does nothing to address climate change
makes solving the problem both harder and more expensive. I am still
optimistic, though, because problems often appear intractable at first. In
the 1980s, as evidence mounted that industrial chemicals were depleting
13

http://www.andrewdessler.com

the ozone, it was not at all clear that we could avoid serious ozone
depletion at a reasonable cost. The chemicals causing the ozone loss,
namely chlorofluorocarbons, played an important role in our everyday life
– in refrigeration, air conditioning, and many industrial processes – just
like the main cause of climate change, fossil fuels, also plays an important
role in our society. But the cleverness of humans prevailed. A substitute
chemical was developed and it seamlessly and cheaply replaced the ozone-
destroying halocarbons – at a cost so low that hardly anyone noticed when
the substitution took place.
Solving the climate change problem will be harder than solving the
ozone depletion problem – how much harder, no one knows. I am
confident, though, that the ingenuity and creativeness of humans is such
that we can solve this problem without damaging our standard of living.
However, there is only one way to find out, and that is to try to do it.
14

Acknowledgments
This book could not have been written without the incredible work of the
climate science community. Ignored by many, demonized by some, I
believe that future generations will look back and say, “They nailed it.” I
hope this book does justice to all of our hard work. The first edition of the
book was written while I was on faculty development leave from Texas
A&M University during Fall 2010. I thank the university for this support.
15

1
An introduction to the climate
problem
◈
We begin our trip through the climate problem by defining weather,
climate, and climate change and by demonstrating how we use latitude and
longitude to describe locations on the Earth. We also discuss something
that few textbooks address: why you should believe this book.
16

1.1 What is climate?
The American Meteorological Society defines climate as
The slowly varying aspects of the atmosphere–hydrosphere–land
surface system. It is typically characterized in terms of suitable
averages of the climate system over periods of a month or more,
taking into consideration the variability in time of these averaged
quantities.
Mark Twain, in contrast, famously summed it up a bit more concisely:
Climate is what you expect; weather is what you get.
Put another way, weather refers to the actual state of the atmosphere at a
particular time. Weather is what we mean when we say that, at 10:53 AM
on November 15, 2014, the temperature in College Station, Texas, was
8°C, the humidity was 66 percent, winds were out of the southeast at 8
knots, the barometric pressure was 30.23 inches, and there was no
precipitation.
Climate, in contrast, is a statistical description of the weather over a
period of time, usually a few decades. It would almost certainly include
average temperature as well as a measure of how much the temperature
varies about this average value, such as the record high and low
temperatures. Figure 1.1 demonstrates one way to look at the climate: It
shows the distribution of daily average temperatures in August near
Fairbanks, Alaska, for two time periods, 1900–1929 and 1970–1999.
During the 1900–1929 period, for example, the most likely daily average
temperature was 10°C, which occurred on approximately 16 percent of the
17

days. Extremes occur less frequently; for example, the probability of
temperatures above 16°C or below 3°C are small. The climate tells us only
the range of probable conditions on a particular day; it contains no
information about what the temperature was on any particular day.
Figure 1.1 Frequency of occurrence of daily average temperature in
August at 64°N, 150°W, near Fairbanks, AK, for two time periods:
1900–1929 and 1970–1999
(data obtained from the twentieth-century reanalysis, version 2,
www.esrl.noaa.gov/psd/data/gridded/data.20thC_ReanV2.html).
In this book, I frequently use the Celsius scale, the standard
temperature scale throughout the world (the Fahrenheit scale more familiar
to U.S. readers is only used in the United States and a few other countries).
For readers who may not be conversant in Celsius, you can convert from
Fahrenheit to Celsius using the equation C = (F – 32) × 5⁄9; or from
Celsius to Fahrenheit, F = C × 9⁄5 + 32. It is also useful to remember that
the freezing and boiling temperatures for water on the Celsius scale are
0°C and 100°C, respectively. On the Fahrenheit scale, these temperatures
18

http://www.esrl.noaa.gov/psd/data/gridded/data.20thC_ReanV2.html

are 32°F and 212°F. Room temperature is about 22°C, which corresponds
to 72°F.
Why do we care about weather and climate? Weather is important for
making short-term decisions. For example, should you take an umbrella
when you leave the house tomorrow? To answer this question, you do not
care at all about the average precipitation for the month, but rather whether
it is going to rain tomorrow. If you are going skiing this weekend, you care
about whether new snow will fall before you arrive at the ski lodge and
what the weather will be while you are there. You do not care how much
snow the lodge gets on average.
Climate, however, is more important for long-term decisions. If you
are looking to build a vacation home, you are interested in finding a place
that frequently has pleasant weather – you are not particularly interested in
the weather on any specific day. Plots like Figure 1.1 can help make these
kinds of climate-related decisions; the plot tells us, for example, that a
house in this location rarely needs air conditioning. If you are building a
ski resort, you want to place it in a location that, on average, gets enough
snow to produce acceptable ski conditions. You do not care if snow is
going to fall on a particular weekend, or even what the total snowfall will
be for a particular year.
An example of the importance of both the climate and the weather
can be found in the planning for D-Day, the invasion of the European
mainland by the Allies during World War II. The invasion required Allied
troops to be transported onto the beaches of Normandy, along with enough
equipment that they could establish and hold a beachhead. As part of this
plan, Allied paratroopers were to be dropped into the French countryside
the night before the beach landing in order to capture strategic towns and
bridges near the landing zone, thus hindering an Axis counterattack.
19

There were important weather requirements for the invasion. The
nighttime paratrooper drop demanded a cloudless night as well as a full
moon so that the paratroopers would be able to land safely and on target,
and then achieve their objectives – all before dawn. The sky had to remain
clear during the next day so that air support could see targets on the
ground. For tanks and other heavy equipment to be brought onshore called
for firm, dry ground, so there could be no heavy rains just prior to the
invasion. Furthermore, the winds could not be too strong because high
winds generate big waves that create problems for both the paratroopers
and the small landing craft that would ferry infantry to the beaches.
Given these and other weather requirements, analysts studied the
climate of the candidate landing zones to find those beaches where the
required weather conditions occurred most frequently. The beaches of
Normandy were ultimately selected in part because of its favorable climate
(tactical considerations obviously also played a key role).
Once the landing location had been selected, the exact date of the
invasion had to be chosen. For this, it would not be the climate that
mattered but rather the weather on a particular day. Operational factors
such as the phase of the tide and the moon provided a window of three
days for a possible invasion: June 5, 6, and 7, 1944. June 5 was initially
chosen, but on June 4, as ships began to head out to sea, bad weather set in
at Normandy, and General Dwight D. Eisenhower made the decision to
delay the invasion. On the morning of June 5, chief meteorologist J. M.
Stagg forecasted a break in the weather, and Eisenhower decided to
proceed. Within hours, an armada of ships set sail for Normandy. That
night, hundreds of aircraft carrying tens of thousands of paratroopers
roared overhead to the Normandy landing zones.
The invasion began just after midnight on June 6, 1944, when British
paratroopers seized a bridge over the Caen Canal. At dawn, 3,500 landing
20

craft hit the beaches. Stagg’s forecast was accurate and the weather was
good, and despite ferocious casualties, the invasion succeeded in placing
an Allied army on the European mainland. This was a pivotal battle of
World War II, marking a turning point in the war. Viewed in this light,
Stagg’s forecast may have been one of the most important in history.
Temperature is the parameter most often associated with climate, and
it is something that directly affects the well-being of the Earth’s
inhabitants. The statistic that most frequently gets discussed is average
temperature, but temperature extremes also matter. For example, it is heat
waves – prolonged periods of excessively hot weather – rather than normal
high temperatures that kill people. In fact, heat-related mortality is the
leading cause of weather-related death in the United States, killing many
more people than cold temperatures do. And the numbers can be
staggering: In August 2003, a severe heat wave in Europe lasting several
weeks killed tens of thousands of people.
Precipitation rivals temperature in its importance to humans, because
human life without fresh water is impossible. As a result, precipitation is
almost always included in any definition of climate. Total annual
precipitation is obviously an important part of the climate of a region.
However, the distribution of this rainfall throughout the year also matters.
Imagine, for example, two regions that get the same total amount of
rainfall each year. One region gets the rain evenly distributed throughout
the year, whereas the other region gets all of the rain in one month,
followed by eleven rain-free months. The environment of these two
regions would be completely different. Where the rain falls continuously
throughout the year, we would expect a green, lush environment. Where
there are long rain-free periods, in contrast, we expect something that
looks more like a desert.
21

Other aspects of precipitation, such as its form (rain versus snow), are
also important. In the U.S. Pacific Northwest, for example, snow that
accumulates in the mountains during the winter melts during the following
summer, thereby providing fresh water to the environment during the
otherwise dry summers. If warming causes wintertime precipitation to fall
as rain rather than snow, then it will run off immediately and not be
available during the following summer. This can lead to water shortages
during the summer.
As these examples show, climate includes many environmental
parameters. What part of the climate matters will vary from person to
person, depending on how each relies on the climate. The farmer, the ski
resort owner, the resident of Seattle, and Dwight D. Eisenhower are all
interested in different meteorological variables, and thus may care about
different aspects of the climate. But make no mistake: We all rely on the
stability of our climate. In particular, food production and freshwater
availability, two of the most important things we rely on to survive, are
greatly affected by the climate. I discuss this in greater depth when I
explore climate impacts in Chapter 9.
A final difference between weather and climate is how easy they are
to determine. Measuring the weather is pretty easy – just walk outside and
look around.1 If you need a higher level of accuracy, you can buy
reasonably cheap instruments to measure the temperature, precipitation, or
any other variable of interest. Climate, in contrast, is much harder to
measure; it requires the gathering of decades of data so that we have
sufficiently good, robust statistics, such as I plotted in Figure 1.1.
22

1.2 What is climate change?
The climate change that is most familiar is the seasonal cycle: the
progression of seasons from summer to fall to winter to spring and back to
summer, during which most non-tropical locations experience significant
temperature variations. Precipitation can also vary by season. In fact,
almost any climate variable can vary over the course of the year.
The concern in the climate change debate – and in this book – is with
long-term climate change. The American Meteorological Society defines
the term climate change as “any systematic change in the long-term
statistics of climate elements (such as temperature, pressure, or winds)
sustained over several decades or longer.” In other words, we can compare
the statistics of the weather for one period against those for another period,
and if the statistics have changed, then we can say that the climate has
changed.
Thus, we are interested in whether today’s climate (defined over the
past few decades) is different from the climate of a century ago, and we
are worried that the climate at the end of the twenty-first century will be
quite different from that of today. To illustrate this, Figure 1.1 shows the
August temperature near Fairbanks, Alaska, for two periods, 1900–1929
and 1970–1999. Clearly, the temperature distributions in these two periods
are different – the temperature distribution at the end of the twentieth
century is about 2°C warmer than at the beginning of the century. In other
words, the climate of this region has changed. It should also be noted that
there is no information on what caused the change – it may be due to
global warming or any number of other physical processes. All we have
identified here is a shift in the climate.
23

The shift in the temperature distribution is only ∼2°C, and it might be
tempting to dismiss this as unimportant. However, as I discuss in Chapter
9, seemingly small changes in climate are associated with significant
impacts on the environment. So you should not dismiss such a change
lightly.
In Chapter 2, we will look more closely at data to determine if the
climate is indeed changing. Before we get to that, however, there are two
things I need to cover. First, in the next section, I discuss the coordinate
system I will be using in this book.
24

1.3 A coordinate system for the Earth
I will be talking a lot in this book about the Earth, so it makes sense to
define the terminology used to identify particular locations and regions on
the Earth.
To begin, the equator is the line on the Earth’s surface that is halfway
between the North and South Poles, and it divides the Earth into a northern
hemisphere and a southern hemisphere. The latitude of a particular
location is the distance in the north-south direction between the location
and the equator, measured in degrees (Figure 1.2). Latitudes for points in
the northern hemisphere have the letter N appended to them, with S
appended to points in the southern hemisphere. Thus, 30°N means a point
on the Earth that is 30° north of the equator, whereas 30°S means the same
distance south of the equator.
Figure 1.2 A schematic plot of latitude.
25

The tropics are conventionally defined as the region from 30°N to
30°S, and this region covers half the surface area of the planet. The mid-
latitudes are usually defined as the region from 30° to 60° in both
hemispheres, and these regions occupy roughly one-third of the surface
area of the planet. The polar regions are typically defined to be 60° to the
pole, and these regions occupy the remaining one-sixth of the surface area
of the planet. The North and South Poles are located at 90°N and 90°S,
respectively.
Latitude gives the north-south location of an object, but to uniquely
identify a spot on the Earth, you need to know the east-west location as
well. That is where longitude comes in (Figure 1.3). Longitude is the angle
in the east or west direction, from the prime meridian, a line that runs from
the North Pole to the South Pole through Greenwich, England, and is
arbitrarily defined to be 0° longitude. Locations to the east of the prime
meridian are in the eastern hemisphere and have the angle appended with
the letter E, whereas locations to the west are in the western hemisphere
and have the letter W appended. In both directions, longitude increases to
180°, where east meets west at the international date line.
26

Figure 1.3 A schematic plot of longitude.
Together, latitude and longitude identify the location of every point
on the planet Earth. For example, my office in the Department of
Atmospheric Sciences of Texas A&M University is located at 30.6178°N,
96.3364°W. Knowing your location can literally be a matter of life and
death – shipwrecks, wars, and other miscellaneous forms of death and
disaster have occurred because people did not know where they were.
Luckily for us, GPS (global positioning system) technology, which is
probably built into your cell phone, can determine your latitude and
longitude to within a few feet.
27

1.4 Why you should believe this
textbook
I now have to address an issue that generally does not come up in college
textbooks: why you should believe it. Students in most classes accept
without question that the textbook is correct. After all, the author is
probably an authority on the subject, the publisher has almost certainly
reviewed the material for accuracy, and the instructor of the class,
someone with knowledge of the field, selected that textbook. Given those
facts, it seems reasonable to simply assume that the information in the
textbook is basically correct.
But climate change is not like every other subject. If you do a quick
Internet search, you will be able to find a Web page that disputes almost
every claim made in this textbook. Your friends and family may not
believe that climate change is a serious problem, or they may even believe
it is a hoax. You may agree with them. This book will challenge many of
these so-called skeptical viewpoints, and you may face the dilemma of
whom to believe.
This situation brings up an important and interesting question: How
do you determine whether or not to believe a scientific claim? If you
happen to know a lot about an issue, you can reach your own conclusions
on the issue. However, no one can be an expert on every subject; for the
majority of issues on which you are not an expert, you need a shortcut.
One type of shortcut is to rely on your firsthand experience about how
the world works. Claims that fit with your own experience are easier to
accept than those that run counter to it. People do this sort of evaluation all
the time, usually unconsciously. Consider, for example, a claim that the
28

Earth’s climate is not changing. In your lifetime, climate has changed very
little, so this seems like a plausible claim. However, a geologist who
knows that dramatic climate shifts are responsible for the wide variety of
rock and fossil deposits found on Earth might regard the idea of a stable
climate as ludicrous, but in turn might be less likely to accept a human
origin for climate change. The problem with relying on firsthand
experience about the climate is that our present situation is unique – people
have never changed the composition of the global atmosphere as much or
as fast as is currently occurring. Thus, whatever the response will be, it
will likely be outside the realm of our and the Earth’s experiences.
Another type of shortcut is to rely on your values: You can accept the
claims that fit with your overall worldview while rejecting the claims that
do not. For example, consider the scientific claim that secondhand smoke
has negative health consequences. If you are a believer in unfettered
freedom, you might choose to simply reject this claim out of hand because
it implies that governments should regulate smoking in public places to
protect public health. Those who are more suspicious of the integrity of
big business are going to be more skeptical of the efficacy of vaccines
because they believe that corporations are willing to put profits ahead of
safety.
Yet another shortcut is to rely on an opinion leader. Opinion leaders
are people who you trust because they appear to be authoritative or
because you agree with them on other issues. They might include a family
member or influential friend, a media figure such as talk show hosts Rush
Limbaugh and Jon Stewart, or an influential politician such as Barack
Obama or George W. Bush. In the absence of a strong opinion of your
own, you can simply adopt the view of your opinion leaders. The problem
with this approach is that there is no guarantee that the opinion leaders
have a firm grasp of the science.
29

The most widely accepted approach is to rely on the opinions of
experts. When the relevant experts on some subject have high confidence
that a scientific claim is true, that is the best indication we have that the
claim actually is true. This is not just my view; I am willing to bet it is
something you believe in, too. If a friend tells you that he thinks he may be
sick, what would you recommend? Your recommendation is likely to be
that he should go see a doctor – and not just any doctor, but one who is an
expert in that particular ailment.
This is also the view of the U.S. legal system. Many court cases
involve questions of science (e.g., what was the cause of death, does a
particular chemical cause cancer, does a DNA sample match the
defendant). To settle those cases, the court will frequently turn to expert
witnesses. These expert witnesses are, as their name suggests, experts on
the matter that they are testifying about, and they provide relevant
expertise to the court to help evaluate the important scientific questions
that a case may revolve around.
To be an expert witness, one must demonstrate expertise in a
particular subject. I have served as an expert witness on climate change in
lawsuits over the permitting of coal-fired power plants, and the court
qualifies me as an “expert” by using my research in climate change as well
as the textbooks I have authored as evidence.
It should be emphasized that one must demonstrate specific, recent
expertise in the exact area under consideration to be an expert witness.
Showing expertise in general technical matters or in a related field is not
sufficient. For example, one might consider a scientist with a Ph.D. in
another field (e.g., solid-state physics) to have a credible opinion about the
science of climate change. This is not so, and a person with a Ph.D. but
without specialized knowledge of the climate would not qualify as an
expert on matters of climate. That also goes for weather forecasters –
30

climate and weather are different, and being an expert in weather does not
qualify someone to be an expert witness on climate. And the requirement
for the expertise to be recent rules out those who were experts, say, a
decade ago but who have not kept up with the latest discoveries in the
field.
There are many more examples that demonstrate that, as individuals
and as a society, we rely on experts when evaluating complex technical
issues. That is probably a good thing, too, because on a planet with 7
billion people, you can always find someone who will contest any claim,
no matter how well established it is. For example, it would be relatively
easy to find someone somewhere who would dispute the claim that
cigarettes cause health problems. So if everyone’s opinion counted equally,
then it would be impossible to ever settle any dispute over a scientific
claim – even one as simple as whether the Earth goes around the Sun.
But we also know that even the most trusted expert can be wrong, so
the opinion of a single expert should be taken with caution. One way to
gain confidence in a particular expert opinion is to ask several experts
instead of just one. We frequently do this for important medical decisions
by getting a second opinion. For high-stakes medical decisions, you would
ideally solicit the opinions of many experts. If all of these experts were to
agree, then you would have justifiably high confidence that the
recommendations are the best advice that modern medicine can provide.
Climate change is really no different. It is obvious that the relevant
experts are the community of climate scientists. And rather than listen to
any single individual, we would do best by asking a large, representative
sample of the world’s climate scientists what they think – and if the vast
majority agree on a particular point, then we can have high confidence that
this is best estimate science can provide.
31

This is, in fact, what has already been done. In 1988, as nations began
to acknowledge the seriousness of the climate problem, the
Intergovernmental Panel on Climate Change (IPCC) was formed. The
IPCC assembles large writing teams of scientific experts and has them
write, as a group, reports detailing what they know about climate change
and how confidently they know it. The reliance on large writing groups
reduces the possibility that the erroneous opinions of an individual or a
small group make it into the report, much like getting multiple opinions in
medicine reduces the chance of a bad diagnosis.
To further minimize the possibility that the group of scientists writing
the report are biased in some direction, the scientists making up the writing
teams are not assembled by a single person or organization; they are
nominated by the world’s governments. Thus, the only way the IPCC’s
writing groups would be biased in some direction is if all of the world’s
governments nominated appropriately biased individuals. This seems
unlikely, particularly since addressing climate change brings a raft of
short-term problems to most governments. Many governments would
therefore be happy if climate change disappeared completely as a political
issue and therefore have no incentive to nominate scientists biased to the
view climate change as a serious problem.
Drafts of the IPCC’s reports are reviewed prior to release by other
expert scientists, and they undergo a public review and a separate review
by the world’s governments. In the end, the IPCC’s reports are widely
regarded as the most authoritative statements of scientific knowledge
about climate change, and as such they carry enormous weight in both the
scientific and the policy communities. In 2007, the IPCC shared the Nobel
Peace Prize in recognition of its work on the climate.
An aside: The Summary for Policymakers
32

If you have ever tried to read an IPCC report, you know that the
1,000 plus page reports can be baffling for non-experts. That is
why every report also has a Summary for Policymakers, a more
readable summary of the full report that runs a few dozen pages.
Referred to as the “SPMs,” they summarize in more general
language the most important conclusions in the main report.
The SPMs also serve another unique function. During a final
meeting after the main report is written, representatives from each
of the world’s governments review a draft SPM written by
scientists and vote on every sentence. Only if there is unanimous
agreement from all of the world’s governments is a sentence
included in the SPM. During this process, sentences are frequently
rewritten to make them acceptable to the world’s governments. If
there is nearly unanimous agreement on a sentence, with just one
or two countries dissenting, than the sentence can be included in
the SPM with a footnote recording the dissent.
The purpose of this exercise is to produce a common set of
scientific facts to serve as the basis of future negotiations on
policy. By having unanimous agreement on every sentence, no
country can later say during policy negotiations that they don’t
agree with a particular scientific fact – they have already agreed to
everything in the SPM.
This means, though, that every country is also trying to mold
the SPM to best suit their negotiating position. During the meeting
for the SPM for the IPCC’s 1995 report, for example, Saudi Arabia
and Kuwait argued strenuously to weaken the statements about
humans causing climate change. When the rest of the world
disagreed, it was then proposed that a footnote would be added to
33

the report noting the disagreement – but the footnote was removed
at Saudi Arabia and Kuwait’s request because it would have been
embarrassing for those two major oil producers to be the only
countries in the world to not accept the scientific evidence of
human impacts on climate.
In the end, the SPM represents a good summary of our
scientific understanding of the climate but one that has an
unavoidable hint of political influence in it. To the extent that
political wrangling affects the SPM, it is almost always to water
down the conclusions – reduce our confidence in scientific
statements, lessen the impacts, etc. But despite these flaws, the
SPMs should be given considerable deference in policy debates
over climate.
Despite the careful process that produces the IPCC reports, it remains
controversial in the public debate. To understand why, consider the curious
case of cigarettes. Scientists have known since the 1950s that smoking
cigarettes is terrible for your health. And in 1964, a landmark report by the
U.S. surgeon general laid out the evidence in great detail for the general
public. The tobacco companies at that point had two choices: they could
accept that their product killed their customers, an admission that would
certainly reduce their profits, or they could fight back by attacking the
science.
Perhaps unsurprisingly, they chose to fight the science. In a memo
released just after the surgeon general’s report, a tobacco executive plotted
out the response: “We must in the near future provide some answers which
will give smokers a psychological crutch and a self rationale to continue
34

smoking. These answers must also point out the weaknesses in the
[Surgeon General’s] Report.”2
Good to their word, the tobacco companies proceeded to wage a
successful multi-decade campaign to cast doubt on the science. It was only
in the 1990s, four decades after the science was actually settled, that the
phony public debate over the health impacts of smoking finally faded
away.
Today, the tobacco debates are the archetype of the dishonest
manipulation of science in pursuit of a particular policy goal.3 And the
dishonesty paid handsomely: it effectively delayed by decades public
awareness of the strength of the scientific consensus of the dangers of
cigarettes, thereby keeping people smoking and delaying government
action to restrict its sales. Given how successful the fight waged against
tobacco science was, it should come as no surprise that those opposed to
political action to address climate change have adopted a similar approach.
Echoing the tobacco debate, one argument frequently made during the
climate debate is that there is, in fact, wide disagreement on the science of
climate change among climate scientists. As evidence, they will point to
Internet petitions and various lists of scientists that dispute the mainstream
view. However, a close evaluation of the dissident scientists on these lists
and petitions reveals that in almost all cases they should not be considered
experts on climate. Although many of the individuals on the lists have
technical degrees, and some even have doctorates, their specific training
does not include climate change. They would never qualify as an expert
witness in a lawsuit on climate change; they do not have the background to
teach a college-level course on the material; and we would never trust the
diagnosis provided by a doctor with equivalent expertise.
The only reason that advocates put such transparently unqualified
people forward as experts is that legitimate experts with the desired
35

opinions are not available. Thus, the lack of credentials of those on the
petitions and lists actually underscores the strong agreement among the
relevant experts on the science of climate change.
A second claim we may hear is that climate scientists are
manufacturing a crisis to benefit themselves. If climate is a crisis, so the
argument goes, then more research funding will flow into the field, the
prestige of climate scientists will increase, and scientists will be able to
implement their preferred social policies.
There is, in fact, no evidence to support this argument. Rather, this
argument relies on the listener simply accepting the obviousness of the
claim that an entire scientific field would be willing to engage in scientific
misconduct for research funding. What is often lost in this discussion is
that all scientific fields have this same incentive. Biomedical fields could
invent a new disease or a cure for an existing one to increase funding;
physicists studying solid-state physics could invent a discovery that could
lead to much faster computers; and space physicists could invent evidence
of life elsewhere in the solar system. All of these “discoveries” would
increase funding and interest in the particular field of interest.
It turns out that there are strong barriers to such widespread fraud by
an entire scientific community. First and foremost, there is a coordination
problem: How do you get everyone to go along with the fraud? The
answer is that you can not. A scientific field such as climate science is a
large, diverse, and intensely competitive endeavor, and the desire to
outthink one’s peers and show that one is smarter than they are is much
greater than the incentive to cooperate in this type of fraud. The reason for
this is that success in science is achieved by impressing one’s colleagues.
One of the best ways to do this is to overturn conventional wisdom, either
by showing that previous scientific results are wrong or by suggesting a
new theory that fits the data better. Because this is so beneficial to the
36

reputation of the individual scientist, it provides a strong incentive against
participating in any conspiracy.
And the incentives in science do not support such a conspiracy.
Money from grants does not generally go into the pockets of the
researchers. Most scientists are employees of federal or state governments
or private universities, and the amount of money they can pay themselves
off grants is extremely limited. Instead, the majority of grant money goes
to buy equipment or pay for graduate students. Thus, research funding
provides a very weak incentive to cheat.
Finally, the entire underlying premise of the “climate science is
corrupt” narrative is questionable. The premise is that, by suggesting that
human effects on the climate are well understood, the field gets more
research funding than it would otherwise. However, history suggests that
whenever a field reaches a conclusion on a problem, funding for further
research on that problem goes down. For example, after ozone depletion
was confidently attributed to ozone-destroying chemicals known as
chlorofluorocarbons in the mid-1990s, the funding for subsequent research
rapidly dropped. By saying that they understand the climate system
reasonably well, members of the climate science community are not
helping their funding. They would do better if they claimed that there was
no consensus on why the climate is warming. In that case, it would almost
certainly be a high priority for most policymakers to fund research into
determining the cause of the warming.
But while there has never been widespread fraud by an entire
scientific community, there have been cases in which advocates opposed
to political action have falsely tried to cast doubt on the science. One such
example was discussed earlier: leaked documents from tobacco companies
clearly show the intent of these companies was to generate doubt in the
public’s mind about the health effects of cigarette smoking.
37

Because of this, we should be leery of the argument that “the experts
can’t be trusted.” It goes against both common sense and our experience in
the real world, and it should only be accepted if extraordinary evidence is
provided. In the climate change debate, such evidence is clearly lacking.
An aside: But I still don’t trust the IPCC
If you talk to people about climate change, you will find some
people who, no matter what arguments you make, will simply not
accept the IPCC reports. To those people, you should point out the
many other reports written by authoritative organizations, such as
the U.S. National Academy of Sciences and the U.K. Royal
Society.4 Or you can look at the statements put out by the scientific
societies that climate experts belong to. In October 2009, for
example, a collection of U.S. scientific organizations sent a letter
to the U.S. Senate stating that climate change is a serious problem
facing the entire human race and that emissions of greenhouse
gases have to be dramatically reduced for us to avoid the most
severe impacts. Signatories of this letter include the American
Association for the Advancement of Science, the American
Chemical Society, the American Geophysical Union, the American
Institute of Biological Sciences, the American Meteorological
Society, the American Society of Agronomy, the American Society
of Plant Biologists, the American Statistical Association, the
Association of Ecosystem Research Centers, the Botanical Society
of America, the Crop Science Society of America, the Ecological
Society of America, the Natural Science Collections Alliance, the
Organization of Biological Field Stations, the Society for Industrial
and Applied Mathematics, the Society of Systematic Biologists,
38

the Soil Science Society of America, and the University
Corporation for Atmospheric Research. Comparable non-U.S.
scientific organizations in other countries have also endorsed the
mainstream view of the science of climate change.
The science in this book follows the reports of the IPCC and the other
relevant national and international scientific organizations. That, in a
nutshell, is why you should believe this book. The alternative views on
climate change you might see or hear, such as those from skeptical friends
or the Internet, do not come from a process as credible as the IPCC’s and
therefore should not have the same standing. It should be emphasized that
this does not mean the IPCC’s reports are correct – any scientific claim is
at risk of being overturned by future research. Nevertheless, the IPCC’s
reports do accurately represent what the relevant experts think about the
science, which is the best guide there is for non-experts.
39

1.5 Chapter summary
Weather refers to the exact state of the atmosphere at a point in
time; climate refers to the statistics of the atmosphere over a period
of time, usually several decades in length or longer.
Climate change refers to a change in the statistics of the
atmosphere over decades. Such statistics include not just the
averages but also the measures of the extremes – how much the
atmosphere can depart from the average.
Temperatures expressed in this book are in degrees Celsius;
conversion from Fahrenheit can be done with this equation: C = (F
− 32) × 5/9.
Any position on the surface of the Earth can be described by a
latitude and longitude. Latitude is a measure of the position in the
North-South direction, while longitude is a measure of the position
in the East-West direction. The tropics cover the region from 30°N
to 30°S; mid-latitudes cover the region from 30° to 60° latitude;
and the polar regions cover from 60° to 90° latitude.
In our society, we frequently rely on experts for advice on highly
specialized or technical fields. For climate change, the IPCC
reports represent the opinion of the world’s experts, and the science
described in this book reflects the IPCC’s scientific views.
Those opposed to policy action on climate change frequently make
doubt of the science of climate change a central part of their
argument. In so doing, they are following the strategy employed by
the tobacco companies to keep public debate about the health
40

effects of smoking alive for decades after the issue was settled by
scientists.
41

Additional reading
The IPCC’s reports are available online from www.ipcc.ch.
A. E. Dessler and E. A. Parson, The Science and Politics of Global
Climate Change: A Guide to the Debate, 2nd ed. (Cambridge: Cambridge
University Press, 2010). Chapter 2 discusses how the scientific and policy
processes work and how assessments (like the IPCC) help bridge the gap
between them.
N. Oreskes and E. M. Conway, Merchants of Doubt: How a Handful of
Scientists Obscured the Truth on Issues from Tobacco Smoke to Global
Warming (London: Bloomsbury Press, 2010). This is an important book
about how deception is used to mislead the public on matters ranging from
the risks of smoking to ozone depletion to the reality of global warming.
Union of Concerned Scientists, Smoke, Mirrors, and Hot Air: How
ExxonMobil Uses Big Tobacco’s Tactics to Manufacture Uncertainty on
Climate Science (Cambridge, MA: Union of Concerned Scientists, January
2007). This is a description of how the tactics employed by the tobacco
companies to cast doubt on the science of the health impacts of smoking
are now being used by oil companies to cast doubt on the science of
climate change (available online at
www.ucsusa.org/assets/documents/global_warming/exxon_report ).
Most scientific societies have statements affirming the science of
climate change presented in this book. This includes statements from the
American Geophysical Union and the American Meteorological Society.
See www.andrewdessler.com/chapter1 for links to the above material
and additional resources for this chapter.
42

http://www.ipcc.ch

http://www.ucsusa.org/assets/documents/global_warming/exxon_report

http://www.andrewdessler.com/chapter1

Terms
Climate
Climate change
Equator
Latitude
Longitude
Mid-latitudes
Opinion leader
Polar region
Prime meridian
Summary for Policymakers
Tropics
Weather
43

Problems
1. Determine the latitude and longitude of the White House, the
Kremlin, the Pyramids of Giza, and the point on the opposite side of
the Earth to where you were born. Use an online tool (e.g., Google
Earth) or an atlas (which you can find in any library).
2.
a) Convert the following temperatures from degrees Fahrenheit to
degrees Celsius: 300, 212, 70, 50, 32, and 0°F
b) Convert the following temperatures from degrees Celsius to
degrees Fahrenheit: 150, 100, 70, 50, 0, and –10°C
3.
a) The temperature increases by 1°C. How much does it increase in
degrees Fahrenheit?
b) The temperature increases by 1°F. How much does it increase in
degrees Celsius?
c) This is true: I told a reporter that the Earth has warmed by 0.8°C
over the last century. When it appeared in print, the sentence said:
Dessler said that the Earth has warmed by 33°F over the last
century. Where did the reporter go wrong?
4. What temperature has a numerical value that is the same in degrees
Celsius as it is in degrees Fahrenheit?
5. Find a two-digit temperature in degrees Fahrenheit for which, if
you reverse the digits, you get that same temperature in degrees
44

Celsius (e.g., find a temperature, such as 32°F, for which the Celsius
equivalent would be 23°C; this example, of course, does not work).
6. Why do you believe that smoking causes cancer? (If you do not
believe this, then why do you believe that smoking does not cause
cancer?) What would be required to get you to adopt the opposing
view?
7. Find two friends who have strong but opposing views of climate
change.
a) Ask both of them why they believe what they do and what
would be required for them to adopt the opposing view. It is
important to understand where their views come from; if they
argue, say, that glaciers are retreating or not, find out where they
get their facts.
b) Which of these positions appears more credible? Why?
c) Can you use their views on climate change to predict their views
on other issues (abortion, gun control) and their political
affiliation?
8. Practice reading a graph. These questions all refer to Figure 1.1.
a) What fraction of days have an average temperature of 15°C
during the 1900–1929 and the 1970–1999 periods?
b) For the 1900–1929 period, what is the warmest temperature that
occurred? What about the 1979–1999 period?
c) What temperature(s) have an equal probability of occurring in
the two periods?
45

d) For the period 1970–1999, estimate the fraction of days that
have a temperature 15°C or greater.
e) Same as d, but for the period 1900–1929.
f) Compare the answers to d and e. What does this tell you about
the changes in extreme heat under even modest warming?
9. Give examples of situations when weather affected your or your
family’s life. Then do the same for climate.
10. Those opposed to the IPCC’s scientific conclusions have set up
their own summary of the science of climate change, which they call
the NIPCC. Do some online research and then compare and contrast
the credibility of the two reports.
11. In the climate debate, few institutions are attacked as frequently
as the IPCC. Using web searches, identify some arguments made by
those arguing that the IPCC cannot be trusted.
1 There are, of course, siting issues in measuring the weather.
Depending on your location, the weather you measure when you walk
outside may not be terribly representative of the weather of the larger
areas.
2 From the Legacy Tobacco Documents Library,
legacy.library.ucsf.edu/tid/ctv74e00/pdf;jsessionid=F68A45A37FAF5E
3AD23A5E84A7EEE463.tobacco03
3 The movie “Thank You for Smoking”
(www.imdb.com/title/tt0427944/) is a great parody of the debate.
4 See the additional reading for this chapter at
www.andrewdessler.com/chapter1 for links to these reports and other
examples.
46

http://legacy.library.ucsf.edu/tid/ctv74e00/pdf;jsessionid=F68A45A37FAF5E3AD23A5E84A7EEE463.tobacco03

http://www.imdb.com/title/tt0427944/

http://www.andrewdessler.com/chapter1

2
Is the climate changing?
◈
In this chapter, I address the question of whether the Earth’s climate is
currently changing and how it has changed in the past. You will see
overwhelming evidence that the climate is indeed changing and that it has
changed significantly over the Earth’s entire history. I will not discuss the
causes of climate change here, though – we will do that in Chapter 7.
In Chapter 1, climate change was defined as a change in the statistics
of the weather over a period of several decades. In this chapter, the statistic
I will primarily focus on is global average temperature for two reasons.
First, the most direct impact from the addition of greenhouse gases to the
atmosphere is an increase in temperature. Changes in other variables, such
as precipitation or sea level, are a response to the temperature change.
Second, we have the best data for temperature. The technology for
measuring it is centuries old, and people have been measuring and
recording the temperature with reasonable global coverage since the
middle of the nineteenth century. In addition to direct temperature
measurements, there are other techniques, such as studying the chemical
composition of ice and rocks, that allow us to indirectly infer the
temperature of the Earth over nearly its entire 4.5-billion-year history.
47

Rather than analyze temperature directly, however, we will instead
analyze temperature anomalies, defined as the difference between the
temperature and a reference temperature; the reference temperature is
usually the average temperature over a previous multi-decadal period. For
example, the monthly average temperature for Australia in August 2009
was 19.4°C, while the August 1979–2009 average for that location was
15.4°C. Thus, that region’s temperature anomaly in August 2009 was
+4.0°C, relative to the 1979–2009 baseline. Had we picked a different
reference period, the anomaly would change. As an example, Figure 2.1
shows the global pattern of anomalies for June 2009.
Figure 2.1 The monthly surface temperature anomaly in June 2009 in
degrees Celsius. The reference temperature for the anomaly is the
average of the June temperatures from 1979 to 2009
(data obtained from the MERRA reanalysis).
Why use anomalies rather than absolute temperature? The main
reason is that absolute temperature can vary sharply over short distances,
such as between a city and a nearby rural area, or between two nearby sites
at different altitudes. You may have noticed this, for example, if your car
displays outside temperature on its dashboard. As you drive a few miles,
48

you might see the temperature change by a few degrees, particularly if you
are driving into and out of a city.
Anomalies, however, are constant over much longer distances: If it is
a degree warmer than average in a city, then it will be a degree warmer
than average a few kilometers away from the city, even if the absolute
temperatures differ by a few degrees. Figure 2.1 shows this – regions of
warm and cold anomalies tend to be hundreds, sometimes even thousands,
of kilometers across. This means that calculations of global average
temperature anomalies require only about a hundred or so temperature
stations spread across the globe. Measuring the absolute temperature of the
planet would require many more stations – more stations, in fact, than
exist.
Another advantage of using anomalies is that you can measure
changes in a quantity even if you cannot measure the absolute value of the
quantity. Imagine, for example, that you want to determine if a child is
growing. The most obvious way to do this is to measure the child’s height
every few months. But, if you could not do that, an alternative would be to
measure his height relative to, say, a mark on the wall. If the top of his
head is 1 inch below the mark one year, even with it the next year, and 1
inch above the following year, then you can be confident that the child is
growing 1 inch per year – even if you never know the child’s absolute
height. This is the situation for the Earth’s temperature. We cannot
measure the Earth’s absolute temperature with high accuracy (because
measuring it would require an extremely dense network of temperature
measurements, as discussed earlier), but we can measure the temperature
anomaly with high accuracy – high enough to see clear warming.
I will also generally focus in our discussion on global average
quantities. The reason is that the climate of a region can vary significantly
just due to weather variability – i.e., particular regions can experience
49

climate extremes (e.g., a heat wave) that are completely independent of
climate change. However, these local variations are usually balanced by an
opposite extreme elsewhere: if one region is undergoing a heat wave, there
is likely another region that is undergoing a cold wave (as seen in Figure
2.1). By averaging over the globe, we rid ourselves of most of this weather
variability and more clearly isolate the smaller climate change signal.
50

2.1 Recent climate change
51

2.1.1 Surface thermometer record
People have been measuring the local air temperature at locations all over
the globe for centuries. They were originally made manually using liquid-
in-glass thermometers, but in recent decades these have been replaced by
automated electronic thermometers. By combining these measurements,
scientists can estimate the global average surface temperature anomaly of
the Earth over the past 150 years, and that time series is plotted in Figure
2.2a.
Figure 2.2 (a) Global annual average temperature anomaly; the gray
line is the annual average, and the black line is a smoothed time series.
(b) Smoothed temperature anomaly time series for three different
regions of the planet: the northern hemisphere (24°N-90°N), the tropics
(24°N-24°S), and the southern hemisphere (24°S-90°S). In both plots,
the reference temperature used in calculating the anomalies is the
1951–1980 average.
Data are from the NASA GISS Surface Temperature Analysis
[Hansen et al., 2010], downloaded from data.giss.nasa.gov/gistemp/.
52

http://data.giss.nasa.gov/gistemp/

The data clearly show that the Earth is warming. From 1880 to 2012,
the average surface temperature of the Earth rose by 0.85°C. The warming
has not been uniform but occurred primarily in two distinct periods, from
1910 to 1945 and from 1976 to 2002. Superimposed on the slow warming
trend are many bumps and wiggles that are unrelated to climate change.
Despite this short-term variability, the recent warming is basically
continuous, with every decade since the mid-twentieth century warmer
than previous decades. The three warmest years in the record were 2005,
2010, and 2014.
It is also worth noting that the year-to-year variations in the global-
average temperature are quite small – just a few tenths of a degree. This is
much smaller than the local temperature variations where you live. Thus,
warming of a few degrees Celsius, which are predicted for the twenty-first
century, would run off the scale in Figure 2.2. While this does not tell us
the effect of such warming, it should certainly compel our attention.
Figure 2.3 shows how the warming of the twentieth century was
distributed across the planet. Clearly, the warming is occurring just about
everywhere – thus justifying the “global” part of “global warming.”
However, the warming has not been entirely uniform. Probably the most
obvious difference is that land areas warmed more than the ocean. Figure
2.2b shows temperatures for the northern hemisphere, tropics, and
southern hemisphere, and it shows that the northern hemisphere warmed
more than the tropics or the southern hemisphere and that the tropics and
southern hemisphere warmed about equally.
53

Figure 2.3 The distribution of warming (in °C) between 1901 and 2012.
Regions where data are too sparse to produce an estimate are white.
Adapted from Figure SPM.1 of IPCC [2013].
In science, no single data set is ever considered definitive, and that is
particularly true of the surface thermometer record. This network of
thermometers was not designed for climate monitoring, and, over the
years, the network has undergone many changes. Changes in the types of
thermometer used, station location and environment, observing practices,
and other sundry alterations all have the capacity to introduce spurious
trends in the data.
For example, imagine you have a thermometer that is in a rural
location in the late nineteenth century. Over time, a nearby city expands so
that by the 1980s, the thermometer is completely surrounded by the city.
Because cities tend to be warmer than nearby rural locations, this would
introduce a warming trend in the data not caused by a warming climate.
Scientists know about these problems, and, to the extent possible,
they adjust the data to take them into account. For example, the impact of a
city growing up around a thermometer can be assessed by comparing the
measurements from that thermometer to nearby thermometers that have
54

remained rural for the entire period. The temperature record in Figure 2.2
includes adjustments to account for as many of these issues as possible.
Nevertheless, uncertainty in the data remains, as a result of both
uncertainties in the adjustments and uncertainties that cannot be adjusted
for. To account for this, scientists put error bars on the trend. An error bar
is the scientists’ estimate of the potential error in their estimate. The name
“error bar” comes from the fact that, on a plot, error bars are frequently
indicated as bars extending from the data. For the warming from 1880 to
2012 of 0.85°C, the error bar is 0.20°C, meaning that the warming is very
likely between 0.65°C and 1.05°C.
Given all of the possible problems in these data, it would be foolish to
rely entirely on this one source to determine if and how much the Earth
was warming. Scientists therefore turn to other data sets to verify this
result. In the rest of Section 2.1, I describe the other data sets used to build
confidence in the surface thermometer data set.
55

2.1.2 Satellite measurements of temperature
It is possible to measure global average temperature from orbit, and the
United States has been flying instruments on satellites to make that
measurement since 1978. Figure 2.4 shows the time series of satellite
measurements of the global monthly average temperature anomaly. These
data show a general warming trend over this period of approximately
0.14°C per decade (1.4°C per century).
Figure 2.4 Satellite measurements of the global monthly average
temperature anomaly (black line). The gray line is temperature from the
surface thermometer record. Anomalies in this plot are relative to the
1981–2010 period
(data obtained from the University of Alabama, Huntsville;
downloaded from
www.nsstc.uah.edu/data/msu/t2lt/uahncdc_lt_5.6.txt).
As with all data sets, though, this one has its own set of problems and
uncertainties. First, satellites actually measure the average temperature of
the lowest 8 km of the atmosphere, from the surface to about the altitude
where airliners fly. Thus, it is not actually a measurement of the surface
56

http://www.nsstc.uah.edu/data/msu/t2lt/uahncdc_lt_5.6.txt

temperature, although the temperature of this layer of the atmosphere
should track the surface temperature.
Another issue with these data is orbital drift of the satellites carrying
the instruments. Imagine that a satellite flies over a location at 2 PM each
day and makes a measurement of that location’s temperature. Over time,
the satellite’s orbit drifts so that it flies over that location later and later
each day. After a few years, the satellite is flying over that location at 3
PM. Because temperatures rise throughout the day, it is generally warmer
at that location at 3 PM than it is at 2 PM. Thus, the drift in the satellite’s
orbit would by itself introduce a warming trend, even if the climate were
not actually changing. This artifact must also be identified and adjusted
for.
Other issues include calibration of the satellite instruments, which
were never designed to make long-term measurements, and the shortness
of the satellite record (just a few decades long), both of which also
introduce uncertainty into the observed warming. As with the surface
thermometer record, these issues are known and adjusted for, to the extent
possible.
One way to gain confidence in the satellite and surface thermometer
records is to compare them; this is done in Figure 2.4 (the same surface
thermometer data, but annually averaged, was plotted in Figure 2.2). The
excellent agreement between these two independent temperature
measurements provides strong confirmation of the reality of the warming
of the climate seen in both data sets.
It is also worth noting that superposed on the long-term warming
trend in Figure 2.4 are lots of shorter-term ups and downs. These are not
random but can be assigned to various physical causes, mainly El Niño-La
Niña cycles and volcanic eruptions. During El Niño events, the Earth
warms several tenths of a degree Celsius. El Niño’s opposite is La Niña,
57

and during those events the Earth cools several tenths of a degree. These
El Niño-La Niña events cause temporary fluctuations in temperature
lasting a few years but no long-term changes in the climate. Volcanic
gases emitted during eruptions cool the climate by blocking sunlight –
after a few years, the effluents are removed from the atmosphere and the
climate returns to normal. These processes will be discussed in more detail
later in the book.
An aside: Has global warming stopped?
“The planet has largely stopped warming over the past 15
years, data shows.”
– FoxNews.com, September 27, 20131
One claim that has arisen recently in the public debate over climate
change is that the Earth was warming, but it stopped warming five,
ten, or fifteen years ago. The implication of this argument is that
climate change is therefore nothing to worry about.
So has climate change stopped? Figure 2.4 showed that
superimposed on the slow warming trend are lots of bumps and
wiggles associated with random variability from things like
weather (a particularly cold winter or a particularly warm summer)
or volcanic eruptions (which cool the planet) or natural cycles like
(El Niño-La Niña cycles).
These sources of short-term variability do not have anything
directly to do with climate change and do not cause any long-term
changes in the climate. The bumps and wiggles do, however, make
determining trends over short time periods (e.g., a decade)
problematic. To illustrate this, Figure 2.5 shows monthly average
58

global surface temperature anomalies between 1970 and 2013.
Over this period, the planet warmed rapidly, at a rate of
1.7°C/century.
Figure 2.5 A plot of monthly and global average surface
temperature from the surface thermometer record (gray line) along
with short-term trend lines (black lines). This figure is an
adaptation of SkepticalScience’s escalator plot
(www.skepticalscience.com/graphics.php?g=47 and
www.skepticalscience.com/still-going-down-the-up-
escalator.html)
Also shown on Figure 2.5 are short-term trends based on
endpoints that were carefully selected to produce cooling trends.
As you can see, it is possible to generate a continuous set of short-
term cooling trends, even as the climate is experiencing a long-
term warming. All you have to do is start the trend calculation
during a particularly hot year (e.g., an El Niño year) and then end it
in a cool year (e.g., a La Niña or volcanic year).
The existence of these short-term negative trends allows
someone to disingenuously claim during almost any year covered
by Figure 2.5 that global warming had stopped or even that the
59

http://www.skepticalscience.com/graphics.php?g=47

http://www.skepticalscience.com/still-going-down-the-up-escalator.html

Earth had entered a cooling period. There is even a term for this
deceptive argument: “Going down the up escalator.”
There are two aspects of this worth emphasizing. First,
claiming that global warming has stopped requires careful
selection of the endpoints. This process of intentionally selecting
data to yield a result counter to the full data set is known as cherry
picking. Many of the skeptical claims you will hear in the public
debate over climate are based on cherry picking a large data set in
order to find the small number of exceptions that support the claim.
Second, it is only possible to find cooling over short time periods.
Over periods lasting several decades, the long-term warming
dominates and even the most egregious endpoint selection cannot
generate a cooling trend. We will return to this point when we talk
about climate predictability in Section 8.7.
So two independent, direct measurements of temperature show the
planet is warming, but this question is so important that even more
confirmation is required. To do this, we turn to other measurements that,
while not direct measurements of temperature, nevertheless tell us
something about the temperature of the planet: the amount of ice on the
planet, the heat content of the ocean, and sea level.
60

2.1.3 Ice
Because ice melts reliably at 0°C, it is a dependable indicator of
temperature. In particular, if the warming trend identified in the surface
thermometer and satellite records is correct, then we should expect to
observe the Earth’s ice disappearing. In this section, I show that ice is
indeed disappearing, thus confirming the warming seen in the other data
sets.
2.1.3.1 Glaciers
Glaciers form in cold regions when snow that falls during the winter does
not completely melt during the subsequent summer. As snow accumulates
over millennia, the snow at the bottom is compacted by the weight of the
overlying snow and turns into ice. This process eventually produces
glaciers hundreds, or even thousands, of feet thick.
The length, areal extent, and total volume of glaciers have been
monitored for decades and, in some cases, centuries. For example, Figure
2.6 shows changes in average glacier length (relative to the length in 1950)
for five world regions over the past few centuries. It shows that glaciers
began retreating around 1800, with the recession accelerating later in the
nineteenth century. The pattern of glacier retreat is consistent worldwide,
confirming that the warming we are now experiencing is truly global.
61

Figure 2.6 Change in mean glacier length over time, measured relative
to length in 1950, for five world regions
(the source is Figure 3.2 of Dessler and Parson, 2010, which was
based on Figure 4.13 of Lemke et al., 2007).
Decreases in precipitation or decreases in cloudiness could also cause
glaciers to recede. However, the fact that glaciers are receding all over the
planet means that, whatever is causing the changes, it must be global. And
there is no evidence of global trends in either cloudiness or precipitation
that could cause the reduction in glacier lengths. We do, however, have
evidence of global trends in temperature (e.g., Figure 2.3). Thus, the
recession of glaciers is consistent with the global warming of the climate
seen in the surface thermometer and satellite records.
2.1.3.2 Sea ice
At the cold temperatures found in polar regions, seawater freezes to form a
layer of ice floating on top of the ocean, typically a few meters thick. The
area covered by sea ice varies over the year, reaching a maximum in late
winter and a minimum in late summer. Given the rapid warming now
occurring, particularly in the Arctic, we would expect to see reductions in
the area covered by sea ice during the summer (during the winter, the
temperatures are so low that a few degrees of warming does not have
62

much of an effect). Figure 2.7 confirms this by showing a clear downward
trend in the area covered by Arctic sea ice at the end of the summer.
Measurements also show that, in addition to shrinking in area, sea ice has
grown thinner.
Figure 2.7 Arctic sea-ice area in September of each year
(data obtained from the National Snow and Ice Data Center; see
nsidc.org/data/seaice_index/).
The Antarctic is a different story. The sea-ice area around that
continent has remained stable since the mid-1970s. This overall pattern –
large losses of sea ice in the Arctic but little loss in the Antarctic – matches
the regional temperature trends in these regions (Figure 2.2b), which
shows large, rapid warming in the northern hemisphere and weaker
warming in the southern hemisphere. In this way, the sea-ice data confirm
not just the overall warming trend but also the global distribution of the
warming.
2.1.3.3 Ice sheets
The Earth has two major ice sheets, one in the northern hemisphere, on
Greenland, and the other in the southern hemisphere, on Antarctica.
63

http://nsidc.org/data/seaice_index/

Although these ice sheets are really just big glaciers, their sheer size puts
them in a class by themselves. They contain the vast majority of the
world’s fresh water and, if they melted completely, the sea level would rise
approximately 65 m. These ice sheets cover millions of square kilometers,
and in places they are more than 3,000 m thick. Figure 2.8 shows the
amount of ice lost between 1992 and 2013 from the Greenland ice sheet.
Over this time, Greenland lost roughly 3,000 billion tons of ice, with the
rate accelerating over time. Measurements from Antarctica show
comparable losses for that ice sheet.
Figure 2.8 Cumulative loss of ice from the Greenland ice sheet, in
billions of tons of ice.
The shaded region is the uncertainty (adapted from Figure 4.15 of
Vaughan et al., 2013).
Putting all the data together, we can conclude that the amount of ice
on the planet is decreasing. This is consistent with measurements of rising
temperatures from surface thermometers and satellites and provides
additional evidence that the Earth is warming.
64

2.1.4 Ocean temperatures
Much of the heat trapped by greenhouse gases goes into heating the
oceans, so we can also look to see if the temperatures of the oceans are
increasing. I am not talking about the surface temperature of the ocean –
that is included in the surface thermometer record already described.
Rather, I am talking about the temperature of the bulk of the ocean: the
water temperature averaged over the entire depth of the ocean (the average
depth of the ocean is 4 km). Scientists determine this temperature by
lowering thermometers into the ocean and measuring the temperature at
various depths and then averaging these results to come up with a single
average ocean temperature over that depth.
Such measurements have been made for several decades, allowing us
to analyze the temperature of the bulk of the ocean since the middle of the
twentieth century; Figure 2.9 plots the time series of the ocean’s
temperature anomaly. The ocean is indeed observed to be warming, and
this provides another source of independent confirmation that the Earth is
warming. While the amount of warming of the ocean appears to be small,
water holds a tremendous amount of energy, so this seemingly small
increase actually represents an enormous accumulation of energy in the
climate system.
65

Figure 2.9 Ocean temperature anomaly in degrees Celsius of the entire
ocean.
Anomalies are calculated relative to the 1970–2000 period (data are
from Balmaseda et al., 2013).
66

2.1.5 Sea level
Sea-level change is connected to climate change in two ways. First, as
grounded ice2 melts, the melt water runs into the ocean, increasing the
total amount of water in the ocean and therefore the sea level. We saw in
Subsection 2.1.3 that we are losing grounded ice all over the planet.
Second, like most things, water expands when it warms, and we saw in
Subsection 2.1.4 that the oceans are indeed warming, and the resulting
expansion also raises sea level.
Figure 2.10 shows that is indeed what is observed. Over the period
1901–2010, global mean sea level rose by 0.19 m based on tide gauge
records and satellite data since 1993. This corresponds to an average rate
of sea level rise of 1.7 mm yr−1. Between 1993 and 2010, the rate was
higher: 3.2 mm yr−1. Thus, sea level is not just rising, the rate of increase
is accelerating.
Figure 2.10 Global-average sea level measured from satellites. The
seasonal cycle has been removed.
Data are described by Nerem et al. (2010) and were downloaded from
sealevel.colorado.edu/content/2013rel7-global-mean-sea-level-time-
67

http://sealevel.colorado.edu/content/2013rel7-global-mean-sea-level-time-series-seasonal-signals-removed

series-seasonal-signals-removed.
Scientists can double-check these values by comparing them to
estimates of the amount of water lost from grounded ice added to the
amount of sea level rise due to warming of the ocean. Although the
calculation is difficult, and there are uncertainties in the measurements
(e.g., the warming of the deepest part of the ocean is not well constrained
by observations), the loss of ice and the warming of the ocean is consistent
with the observed change in sea level. Such detailed comparisons increase
our confidence that the changes in each data set are accurate.
68

2.1.6 Putting it all together: Is today’s climate changing?
The answer is an emphatic yes! In fact, the evidence is so strong that the
IPCC calls today’s warming “unequivocal” – meaning it is beyond doubt.
It is worth exploring the source of high confidence in this conclusion.
First, there is great consistency among the various data sets. The surface
thermometer record and the satellite record agree well, and both show that
temperatures are rising. The loss of ice on the Earth’s surface is consistent
with these increasing temperatures, as is the increase in the heat content of
the ocean. Finally, the observation of increasing sea level fits with both the
loss of ice and the increasing heat content of the ocean.
These data sets are fundamentally independent. For example, issues
such as changes in the station environment, which may affect the surface
thermometer record, do not affect the satellite record. Issues such as orbit
drift affect the satellite record but do not affect the surface thermometer
record. And neither of these problems affects the measurements of glacier
length or sea level. This means that there is no single problem or error that
could introduce a spurious warming trend in all of the data. Because of
this, there is virtually no chance that enough of these data sets could be
wrong by far enough, and all in the same direction, that the overall
conclusion that the climate is currently warming is wrong.
Moreover, the data sources we have reviewed are just a small part of
the mountain of evidence that the Earth is warming. Other corroborating
evidence includes decreased northern hemisphere snow cover, thawing of
Arctic permafrost, strengthening of mid-latitude westerly winds, fewer
extreme cold events and more extreme hot events, increased extreme
precipitation events, shorter winter ice seasons on lakes, and thousands of
observed biological and ecological changes that are consistent with
69

warming (e.g., poleward expansion of species ranges and earlier spring
flowering and insect emergence). Not every data set shows warming, but
such contrary data are rare, regionally limited, and vastly outnumbered by
evidence of warming.
70

2.1.7 What is not evidence of climate change
It is useful at this point to recognize what is not evidence of climate
change. Because climate change is a shift in the statistics of the
atmosphere, a single seemingly odd weather event is almost never
evidence of climate change. A single extremely hot summer, for example,
even if it were hotter than any other summer of the past 100 years, might
nonetheless occur in a stable climate. If hot summers were to begin to
happen regularly, however, then that would be indicative of climate
change.
It is also important to avoid drawing conclusions about the global
climate from regional climate extremes. Figure 2.1, for example, shows
that temperatures in South America were much colder than average in June
2009. People living there might be forgiven for wondering where global
warming was, but if they concluded as a result of those regional anomalies
that climate change was no longer happening, they would be wrong. In that
case, other regions were hotter than average, and those compensated for
the low temperatures found in South America.
So be careful when evaluating the evidence for and against climate
change. Do not be misled by unlikely-seeming single events or by regional
occurrences. Neither is indicative of a shift in the global climate.
71

2.2 Climate over the Earth’s history
72

2.2.1 Paleoproxies
To put today’s warming into context, it is useful to consider the Earth’s
entire climate history. The measurements described in the previous
section, however, go back at most a few centuries, so other data sets are
needed if we wish to look back any further. Such data sets are known as
paleoproxies, which are long-lived, geological, chemical, or biological
systems that have the climate imprinted on them. In this way, we can make
measurements today that tell us what the climate was like in the past.
For example, the ice in a glacier or ice sheet provides useful climate
data dating back to the time when the snow fell. Remember that glaciers
and ice sheets form when snow accumulates from one year to the next and
is converted to ice by the weight of the overlying snow. The chemical
composition of the ice holds important information about the air
temperature around the glacier when the snow fell, as do variations in the
size and orientation of the ice crystals. Small air bubbles trapped during
the formation of glacial ice preserve a snapshot of the chemical
composition of the atmosphere when the snow fell. In addition, the dust
trapped in the ice gives information about prevailing wind speed and
direction. And because more dust blows around during droughts, it also
provides information about how wet or dry the regional climate was when
the ice formed. Finally, because sulfur is one of the main effluents of
volcanoes, measurements of sulfur in glacial ice show whether there was a
major volcanic eruption around the time the ice formed.
To obtain all of this information, ice cores are obtained by drilling
down into the glacier or ice sheet with a hollow drill bit and extracting a
cylinder of ice a few inches in diameter. Reconstructing past climate
information from an ice core then requires two steps. First, the age of each
ice layer must be determined from its depth inside the glacier. The deeper
73

down the ice was obtained, the older the ice is and the further back in time
for which it provides climate information. Much effort has been spent
connecting a particular chunk of ice to an exact time, because the rate of
ice accumulation varies over time and because ice inside a glacier can
compress and flow under the weight of the ice above. Second, the
characteristics actually observed, such as the abundance of chemicals in
the ice, must be translated into the climatic characteristics of interest, such
as temperature.
Obviously, ice cores only provide climatic information in regions and
over time periods that are cold enough for permanent ice to exist. This
includes Greenland, Antarctica, and glaciers found around the world. Ice
cores from the thickest, oldest ice in Antarctica have provided climate
reconstructions dating back an amazing 800,000 years.
However, there are other paleoproxies that provide data in other
regions and over other time periods. For example, trees also store climate
information in their tree rings. Tree growth follows an annual cycle, which
is imprinted in the rings in their trunks. As trees grow rapidly in the spring,
they produce light-colored wood; as their growth slows in the fall, they
produce dark wood. Because trees grow more, and produce wider rings, in
warm and wet years, the width of each ring gives information about
climate conditions around that tree in that year. So by looking at the rings
of a tree, scientists obtain an estimate of the local climate around the tree
for each year during which the tree was alive.
Climate data from tree rings are only available for a fraction of the
Earth’s surface. They are obviously not available for oceans, or for desert
or mountainous areas where no trees grow. They are also not available in
the tropics, where the weaker seasonal cycle causes trees to grow year
round; those trees do not produce rings. Finally, tree rings only reveal
information about the climate as far back as trees are available. This means
74

that the longest tree ring records provide useful climate information for the
past millennium or so.
Ocean sediments, which accumulate at the bottom of the ocean every
year, also contain information about climate conditions at the time they
were deposited. The most important source of information in sediments
comes from the skeletons of tiny marine organisms. The relative
abundance of species that thrive in warmer versus colder waters gives
information about surface water temperature. The chemical composition of
the skeletons and variations in the size and shape of particular species
provide additional clues. In total, ocean sediments provide information
about water temperature, salinity, dissolved oxygen, atmospheric carbon
dioxide, nearby continental precipitation, the strength and direction of the
prevailing winds, and nutrient availability; this information goes back tens
of millions of years.
Putting all of these paleoproxies together gives us a reasonably
complete picture of the global climate going back many millions of years,
with some information about the climate going back billions of years.
75

2.2.2 The Earth’s long-term climate record
Although many of details of the climate during the first 97 percent of the
Earth’s history are unknown, there are a few things that we can say. To
begin with, the oldest sedimentary rocks on the planet are nearly 4 billion
years old. Because sedimentary rocks generally form in the presence of
liquid water, their existence suggests that the Earth has been warm enough
over most of its history that water has remained mostly in the liquid phase.
While the Earth has generally been warm, there is also evidence of
intervals of widespread ice cover (known as a glaciations). The evidence
comes in the form of marks on rocks, such as abraded rock surfaces and
other geologic formations that form when giant ice sheets flow over rocks.
It shows that, approximately 700 million years ago, the Earth was covered
by ice from the poles to near the equator – a climate configuration now
referred to as snowball Earth. There was also a significant planetary
glaciation roughly 300 million years ago.
Figure 2.11 shows a reconstruction derived from ocean sediments of
polar temperatures over the past 70 million years. The warmest
temperatures in this record occurred approximately 50 million years ago –
15 million years after the extinction of the dinosaurs – in a period called
the Eocene Climatic Optimum. During that time, the planet was far warmer
than it is today. Forests covered the Earth from pole to pole, and plants that
cannot tolerate even occasional freezing lived in the Arctic, along with
animals such as alligators that today live only in tropical climates. Since
that time, the Earth has experienced a long-term cooling. Clearly, humans
had nothing to do with either the warmth of the Eocene or the cooling
since then; I will talk more about the causes of these climate variations in
Chapter 7.
76

Figure 2.11 Reconstructed temperature of the polar regions over the
past 70 million years. The sharp temperature spike 55 million years ago
represents the PETM. The gray bar on the right shows the past 4 million
years, which are expanded in Figure 2.12. Because this time series is
from ocean isotopes, it is sensitive both to temperature and to the total
volume of ice on land. Starting roughly 35 million years ago, some of
the variation here comes from changes in land ice volume rather than
temperature. The overall trend, however, mostly represents changes in
temperature
(the source is Figure 3.8 of Dessler and Parson, 2010, which was
adapted from Figure 2 of Zachos et al., 2001).
Figure 2.11 also shows the Paleocene-Eocene Thermal Maximum
(frequently abbreviated PETM), which occurred 55 million years ago, at
the temporal boundary between the Paleocene and Eocene epochs. The
PETM featured an abrupt warming of a few degrees Celsius that occurred
over a few thousand years. The temperature then slowly returned back to
pre-PETM temperatures over the next few hundred thousand years. It is
believed that this was caused by a massive release of greenhouse gases;
many scientists view this episode as a good analog to the warming event
we are now in the midst of. I will also discuss this event in more detail in
Chapter 7.
77

Figure 2.12 zooms in to show global temperature variations over the
past 4 million years. Like the 70-million year record in Figure 2.11, this
record also shows a general cooling trend. This record, however, covering
a more recent time, shows fine-scale details that are not visible in the
longer record. For example, starting roughly 3 million years ago, around
the time that large ice sheets first appeared in the northern hemisphere,
large oscillations between warmer and cooler periods suddenly appear in
the record. During the cool periods, called ice ages, the ice sheets
expanded to cover large parts of the northern hemisphere’s land areas.
During the warm periods between the ice ages, called interglacials, the ice
sheets contracted. From approximately 2.5 million to 1 million years ago,
ice ages occurred every 41,000 years. Since then, for reasons that are not
well understood, the frequency of ice ages shifted to every 100,000 years
and the magnitude of the ice-age cycles increased.
Figure 2.12 Measurement of global average relative temperature over
the past 4 million years. The vertical axis measures the relative
abundance of oxygen-18, the heavy isotope of oxygen that is a proxy for
temperature, in ocean sediment cores. The temperature difference
between the top and bottom of the graph is roughly 10°C. The gray bar
on the right shows the past 410,000 years, which are expanded in Figure
2.13
(the source is Figure 3.9 of Dessler and Parson, 2010, which is based
on the analysis of Lisiecki and Raymo, 2005).
78

Figure 2.13 zooms in again, showing a record of temperature and
carbon dioxide levels for the Antarctic region over the past 410,000 years
constructed from ice cores. This record shows in more detail the shape of
ice-age cycles – the cooling into an ice age is slow, taking several tens of
thousands of years, whereas the warming at the end of an ice age occurs
faster, in approximately 10,000 years. Overall, these ice ages lasted about
100,000 years, while the interglacials are relatively short, lasting 10,000 to
30,000 years. The last ice age ended roughly 12,000 years ago, and since
then the Earth has been enjoying a rather pleasant interglacial.
Figure 2.13 Temperature anomaly of the southern polar region (solid
line) over the past 410,000 years, relative to today’s temperature,
constructed from an Antarctic ice core. Carbon dioxide (dotted line) is
from air bubbles trapped in the ice
(the source is Figure 3.10 of Dessler and Parson, 2010, which was
adapted from Petit et al., 1999).
It is particularly noteworthy that atmospheric carbon dioxide, which
can also be estimated from the ice core, varies closely with atmospheric
temperature over these ice age cycles. I will discuss the implications of
this relationship in detail in Chapter 7, but you would be correct if you
concluded that carbon dioxide variations play a key role in the generation
of the ice-age cycles.
79

Figure 2.14 zooms in again to show the global temperature of the
Holocene, the period beginning at the end of the last ice age, about 11,700
years ago. This estimate shows that temperatures peaked about 7,000 years
ago and then started a slow, long-term decline that bottomed out in a
period 200 to 300 years ago, known as the Little Ice Age, during which
temperatures were about 1°C below today’s. After the Little Ice Age,
temperatures began warming. Today’s temperatures are certainly warmer
than most of the Holocene; however, given all of the uncertainties, it is
unclear whether the mid-Holocene was warmer than today.
Figure 2.14 Global temperature anomaly of the last 11,000 years, based
on multiple proxy records. The shaded represents the uncertainty in the
estimate.
Anomalies are calculated relative to the 1961–1990 average (the
source is Figure 1B of Marcott et al., 2013).
An aside: What does the paleorecord tell us about how serious of a
threat climate change is?
80

As we will talk about in Chapter 8, forecasts for the twenty-first
century are for a few degrees Celsius of warming. That might not
seem like much, but the paleorecord says otherwise. The ice ages
were only about 5°C colder than today, and the Earth was
essentially a different planet. Glaciers several thousand feet thick
covered much of North America, sea level was 100 m lower than
today, and there were significant accompanying changes in the
world’s environment and ecosystems.
The Little Ice Age was only about 1°C below today’s
temperature, a seemingly trivial amount, but the climate was
different enough that we call it the Little Ice Age. Glaciers in
Europe advanced dramatically, destroying numerous farms and
villages. Paintings from the time show a cold and snowy climate
that does not exist today. In London, the freezing of the Thames
River, commonplace during the era, was celebrated with a winter
fair that took place on the frozen river – that river no longer freezes
over. In their camp in Valley Forge, PA, the Continental Army
nearly froze to death during the winter of 1777–1778.
Thus, we should expect warming of a few degrees Celsius, if
it comes to pass, to radically change the planet we’re living on.
Figure 2.15 zooms in one last time to show average northern
hemisphere temperature over the past 1,000 years, based on multiple
proxies and modern records. Once again, this figure shows short time-scale
temperature variations that are not visible in the graphs covering longer
time spans. The various estimates differ, particularly before Year 1500, but
all show a similar pattern. Temperatures were warm 1,000 years ago,
during a period known as the Medieval Warm Period. There were then
81

several centuries of gradual cooling, bottoming out in the Little Ice
Age,followed by faster warming since the nineteenth century.
Figure 2.15 Average northern hemisphere temperature anomaly over
the past 1,000 years, based on multiple proxy records and the modern
surface thermometer record.
Anomalies are calculated relative to the 1961–1990 average (the
source is Figure 3.11 of Dessler and Parson, 2010, which was adapted
from Figure S-1 of the National Research Council, 2006).
This vast and growing body of information about the Earth’s past
allows us to make several important conclusions. We can say with high
confidence that the past few decades of the twentieth century are warmer
than any comparable period during the past 400 years, and possibly even
warmer than the peak of the Medieval Warm Period, around 1,000 years
ago. It is presently unclear whether today’s temperatures exceed the
previous peak temperature of the Holocene, which occurred about 7,000
years ago.
Over geologic time scales (millions of years or longer), we can say
with high confidence that the Earth has been both far warmer and far
cooler than it is today. But this fact provides no information about the
cause of today’s warming. As I will discuss in Chapter 7, there is strong
82

evidence to suggest that the recent warming is primarily due to human
activities.
We can also say that the warming of the past few decades has been
rapid. For example, the warming over the past century (approximately
0.8°C in 100 years) is occurring at least ten times faster than the average
rate of warming coming out of the last ice age (roughly 5°C in 10,000
years corresponds to an average warming of 0.05°C/century). This means
that projections of warming of several degrees Celsius over the next
century would be both large and exceptionally rapid.
The challenge for the scientific community is to come up with a
theory that explains all of the variations in the climate record, from
snowball Earth 700 million years ago to the rapid warming of the past few
decades. In the next few chapters, we will learn about the fundamental
physics that governs our climate, and then, in Chapter 7, we will put it all
together to show that most of the warming of the past few decades can be
attributed to human activity.
83

2.3 Chapter summary
The most well-studied and reliable source of temperature data for
the past century is the surface thermometer network. It shows a
global and annual average warming of 0.85°C between 1880 and
2012, with an uncertainty of ±0.2°C.
Scientists have a large number of independent measurements with
which to confirm the warming seen by the surface thermometer
network. These include satellite measurements of temperature,
measurements of the amount of ice on the planet, ocean heat
measurements, and sea-level measurements. All of these data
confirm the warming seen in the surface thermometer data.
Because of the overwhelming evidence supporting it, the scientific
community has concluded that the observed warming of the
climate system is beyond doubt – the IPCC uses the word
“unequivocal.” Furthermore, scientists have concluded that the
previous decade is very likely the hottest of the past 400 years.
Looking back further in time, we see that the Earth’s climate has
varied widely over its 4.5-billion-year history. The geologic record
shows that the climate has been both warmer and cooler than
today’s climate.
Over the past few million years, the Earth has oscillated between
ice ages and warmer interglacial periods. Ice ages are about 5°C
cooler than the interglacials. The Earth is currently in an
interglacial.
84

Additional reading
The Working Group I report of the IPCC’s Fifth Assessment describes, at
varying levels of detail, the evidence supporting the conclusion that
warming is “unequivocal.” For the most detail, see Chapter 2 of the report.
For a less detailed overview, see Section TS.2 of the Technical Summary.
And, for a short, high-level discussion, see Section B of the IPCC’s
Summary for Policymakers (you can download all of these at
www.ipcc.ch/report/ar5/wg1/).
One of the internet’s great resources for climate information is
Skeptical-Science.com. It has many articles about the science of climate
change, including discussions of the quality of the various temperature
records. For example, here is a nice article on the reliability of the surface
thermometer measurements: www.skepticalscience.com/surface-
temperature-measurements.htm. When you are confronted with a claim
that sounds wrong (“We’re entering a new ice age!”), this is the first place
you should go to check it out.
See www.andrewdessler.com/chapter2 for additional resources for
this chapter.
85

http://www.ipcc.ch/report/ar5/wg1/

http://www.skepticalscience.com/surface-temperature-measurements.htm

http://www.andrewdessler.com/chapter2

Terms
Cherry picking
Eocene Climatic Optimum
Error bar
Ice ages
Ice core
Interglacials
Little Ice Age
Medieval Warm Period
Paleoproxies
Snowball Earth
Temperature anomaly
86

Problems
1. Every year, you measure the height of a child relative to a coat
hook on the wall. In the first year, he was 2″ below the hook, the next
year he was 1.5″ below the hook, the next year he was 0.75″ below
the hook, the next year he was even with the hook, the next year he
was 0.5″ above the hook, and the last year he was 1.5″ above the
hook. (a) What was the total amount the child grew? (b) What was his
average growth rate (in inches per year)? (c) What was his absolute
height at the end of the last year?
2. How much did the Earth warm between a) 1880 and 2012 and b)
1970 and 2012? Provide answers in both degrees Celsius and degrees
Fahrenheit.
3. If you found out that the satellite data were unreliable because of a
previously unknown error, would that change your opinion about
whether the Earth is currently warming? Why or why not?
4. A reporter asks you to explain why scientists are so confident that
the Earth has undergone a general warming over the past few
decades. Knowing that reporters hate long answers, construct an
answer that takes 60 seconds or less to deliver.
5. List the evidence that supports the contention that the Earth is
currently warming. Is there any evidence that goes against this
conclusion?
6. What is a temperature anomaly? Why are temperature anomalies
typically used in global temperature calculations?
87

7. Download the annual and global average temperature data from the
NASA GISS (Google the term GISTEMP) and reproduce Figure 2.2.
Calculate your own trends for the past thirty years and for the past
100 years.
8. For the more adventurous students: Download the monthly and
global average temperature data from the NASA GISS (Google the
term GISTEMP). Calculate trends over various lengths of time
(ranging from a decade to several decades). Over what length of time
can you frequently find negative trends; over what length of time are
the trends mainly positive?
9. Why do we turn to paleoproxy measurements to infer the
temperature of millions of years ago?
10. Go to cdiac.ornl.gov/epubs/ndp/ushcn/access.html and plot up the
monthly temperature for the past 100 years for the station nearest
your hometown. Does this look like the global average time series in
Figure 2.2? Should it?
11. A global warming advocate tells you that the Earth is now warmer
than it has ever been. Is that correct?
1 www.foxnews.com/science/2013/09/27/un-climate-change-report-
dismisses-slowdown-in-global-warming/, accessed 11/15/14.
2 Grounded ice is ice that is resting on land. When it melts and the water
runs into the ocean, sea level rises. This is different from floating ice.
When floating ice melts, the melt water occupies roughly the same
volume as was displaced by the ice (as shown by Archimedes), so there
is little sea level rise. Thus, melting glaciers and ice sheets cause sea
level to rise, but melting sea ice causes nearly none. I say “nearly”
because sea ice is pure water, while seawater has salt in it; the
88

http://cdiac.ornl.gov/epubs/ndp/ushcn/access.html

http://www.foxnews.com/science/2013/09/27/un-climate-change-report-dismisses-slowdown-in-global-warming/

differences in density between pure water and salt water leads to a small
amount of sea level change when sea ice melts.
89

3
Radiation and energy balance
◈
The Earth’s climate is a complex physical system. Nevertheless, we can
still understand much about the climate even without an advanced degree
in physics. In this chapter, I introduce the important physics required to
understand the climate. Then, in Chapter 4, we will use this physics to
construct a simple model of our climate.
90

3.1 Temperature and energy
Before we get into the physics of climate, it is useful to first talk about the
concept of energy. To a physicist, energy is the capacity to do work – such
as lifting a weight, turning a wheel, or compressing a spring. The unit of
energy most frequently used in physics is the joule, abbreviated as the
letter J, and 1 J is approximately the amount of energy required to lift 100
g about 1 m – or to lift an apple about 3 ft.
Energy often moves from one place to another. The rate at which
energy is flowing is referred to as power. It is usually expressed in watts,
abbreviated as the letter W. One watt is equal to one joule per second – that
is, 1 W = 1 J/s – so a 60-W light bulb consumes 60 J of energy every
second.
An analogy may help to illuminate the difference between power and
energy. A gallon is a quantity, such as a gallon of water. This is akin to a
joule, which is a quantity of energy. The rate at which water flows through
a pipe is measured in, say, gallons per minute. The rate at which energy
flows is the power, and it is measured in watts (joules per second). As you
will see in this chapter and the next, climate is all about energy flows, so
climate calculations mostly focus on power.
An example: How much power does it take to run a human body?
A typical human consumes approximately 2,000 food calories per
day. Calories are an alternative unit of energy, where 1 food
calorie = 4,184 J. Thus, 2,000 food calories corresponds to
8,368,000 J = 8.37 MJ. One day has 24 hours × 60 minutes × 60
91

seconds = 86,400 seconds in it, so dividing 8,368,000 J by 86,400 s
yields 97 J/s = 97 W. Thus, the typical human requires roughly 100
W to power his or her body – about the same power required to run
a light bulb or two. One horsepower is approximately 740 W, so
another way to think about this is that it takes about one seventh of
a horsepower to run your body.
The internal energy of an object refers to how fast the atoms and
molecules in the object are moving. In a cup of water, for example, if the
water molecules are moving slowly, then the water has less internal energy
than another cup in which the molecules of water are moving rapidly. In a
solid, the movements of the atoms are approximately fixed in space by
intermolecular forces – that is why it is a solid. The atoms, however, can
still move small distances around their fixed position. The faster these
atoms move about their fixed position, the more internal energy the object
has.
This brings us to a concept that most people are familiar with:
temperature. Temperature is a measure of the internal energy of an object.
As an object’s internal energy increases and the molecules of the object
speed up, the temperature of the object also increases. Thus, if you have
two cups of water, one hot and the other cold, you can conclude that the
water molecules in the hot cup are moving faster than the water molecules
in the cup of cold water.
In Chapter 1, I introduced the Celsius temperature scale, which
scientists use frequently. There is another temperature that is even more
favored by physicists, and it is called the Kelvin scale. The temperature in
Kelvin is equal to the temperature in degrees Celsius plus 273.15 (K = C +
273.15). Thus, the freezing temperature, 0°C, is equal to 273.15 K,
92

whereas the boiling temperature, 100°C, is equal to 373.15 K. “Room
temperature” is 22°C or so, which is 295 K. Most temperatures found in
the Earth’s atmosphere are between 200 K and 300 K, and the average
surface temperature of the Earth (today, at least) is roughly 288 K.
Physicists prefer the Kelvin scale because temperature expressed in
Kelvin is proportional to internal energy. Thus, if the temperature doubles
from 200 K to 400 K, then the internal energy of the object also doubles. If
the internal energy of an object increases by 10 percent, then the
temperature expressed in Kelvin also increases by 10 percent. And 0 K is
absolute zero – the temperature at which molecules have zero internal
energy and cease moving1; this is the coldest possible temperature.
Because of this important quality, the physics equations introduced in this
chapter and the next require energy to be expressed in Kelvin.
93

3.2 Electromagnetic radiation
It has long been recognized that the warmth of our climate is provided by
the Sun. However, the Sun sits 150 million km away from the Earth, with
the vacuum of space in between. How does energy from the Sun reach the
Earth?
Energy is transported from the Sun to the Earth by electromagnetic
radiation.2 Electromagnetic radiation includes visible light, like that put
out by your desk lamp or the Sun, X-rays, like those that allow us to detect
broken bones, microwaves, like those that cook your dinner, and radio-
frequency waves, like those that bring calls to your cell phone and WiFi to
your computer.
One can think of electromagnetic radiation as a stream of photons,
small discrete packages of energy.3 As photons travel from Point A to
Point B – such as from the Sun to the Earth – each one carries a small
amount of energy, and this is how energy is transported from the Sun to
the Earth.
Photons have a characteristic size, referred to as the wavelength,
which determines how the photons interact with the world. Photons with
wavelengths of between 0.3 and 0.8 microns (a micron, abbreviated as µm,
is a millionth of a meter; a human hair is 100 µm or so in diameter) can be
seen with the human eye – so we refer to these photons as visible. Within
the visible range, the different wavelengths appear to the human eye as
different colors (Figure 3.1). Humans see photons with wavelengths near
0.4 µm as blue, 0.6 µm as yellow, and 0.8 µm as red.
94

Figure 3.1 The electromagnetic spectrum. Note that the visible part
makes up only a minor part of this spectrum.
Photons with longer wavelengths, from 0.8 to 1,000 µm, are termed
infrared – from the Latin for “below red” – because they are beyond the
red end of the visible spectrum. Despite being invisible to humans, these
photons play an important role in both the Earth’s climate and in our
everyday lives. Photons with wavelengths just below the human detection
limit of 0.3 µm are called ultraviolet because their wavelength is beyond
the violet end of the visible spectrum.
Photons with wavelengths between 1,000 µm (1 mm) and 0.3 m are
termed microwaves, and photons in this wavelength range are used in
many familiar applications, from cooking to radar. Wavelengths longer
than about 0.3 m are radio-frequency waves, and they are used, as the
name implies, in radio applications. The entire electromagnetic spectrum is
diagrammed in Figure 3.1.
The wavelength determines a photon’s physical properties. For
example, visible and infrared photons cannot go through walls, but radio-
frequency photons can. The human eye can detect visible photons but not
infrared or microwave photons. When you get a full body scan at the
airport’s security checkpoint, the machine is most likely using microwaves
– that wavelength goes through clothes but is stopped by denser materials
such as flesh, a bomb, or a gun. Finally, the atmosphere is transparent to
95

visible photons but not to infrared photons; this fact has enormous
implications for our climate and will be discussed at length in Chapter 4.
96

3.3 Blackbody radiation
We know that both the Sun and the lamp on your desk are emitting
photons. After all, you can see the visible photons that they are emitting.
They are not, however, the only things around you that are emitting
photons. In fact, everything around you is emitting photons all of the time.
So right now, you are emitting photons, as are the walls of the room you
are sitting in, your desk, your dog, this book. Everything.
If everything is emitting photons, then why does not everything glow
like a light bulb? It turns out that the wavelength emitted is determined by
the object’s temperature. Figure 3.2 plots emissions spectra for idealized
objects called blackbodies at three temperatures. An emissions spectrum
shows the power carried away from an object by the photons at each
wavelength.
97

Figure 3.2 Power emitted at different wavelengths from objects (with
surface area of 1 m2) at three temperatures: (a) 300 K, (b) 1,600 K, and
(c) 6,000 K. The vertical axes are in units of 1 W/µm, 1,000 W/µm, and
1 MW/µm of wavelength range, respectively. Gray bars show the
wavelength range visible to human eyes.
Figure 3.2a shows the distribution of photons emitted by a 300-K
blackbody, about room temperature. Photons emitted by this object almost
exclusively have wavelengths greater than 4 µm or so. These wavelengths
are outside the range that is visible to humans (indicated by the gray
98

(3.1)
shading in the figure). Thus, all room-temperature objects are emitting
photons, but you cannot see the photons because they fall outside the
visible range. This is, in fact, the origin of the term blackbody. At room
temperature, the object appears black because the photons emitted by these
objects are invisible to humans (also playing a role is that blackbodies
absorb all photons that fall on them – they do not reflect any). Blackbodies
are idealized constructs, but most objects nevertheless behave like one, at
least approximately.
Figure 3.2a also shows that the peak of the emissions spectrum for a
300-K blackbody occurs near 10 µm. It turns out that there is a simple
relation between the temperature and the peak of the object’s emission
spectrum. This relation is known as Wien’s displacement law:
T is the temperature of the blackbody in Kelvin and λmax is the
wavelength of the peak of the emission spectrum in microns. If we put 300
K into Equation 3.1, we get 9.7 μm, which is in good agreement with
Figure 3.2. Note the importance of using Kelvin temperature – had I used
the temperature in degrees Celsius, I would have calculated λmax =
2897/27 = 107 μm, which would be incredibly wrong.
Wien’s displacement law also tells us that, as an object heats up, λmax
decreases, shifting the peak of its emission spectrum to shorter
wavelengths. Thus, a 1,600-K object has λmax = 1.9 μm and a 6,000-K
object has λmax = 0.5 μm, values consistent with the emissions spectra in
Figures 3.2b and 3.2c.
It is worth emphasizing that objects do not just emit photons at λmax;
they emit them over a range of wavelengths around λmax. So, while λmax =
1.9 μm for the 1,600-K object, the object emits photons over a range of
wavelengths from 0.7 to 10 µm. Because a fraction of the photons emitted
99

by this object have wavelengths between 0.7 and 0.8 µm, which lie at the
red end of the visible spectrum, humans will perceive a 1,600-K object as
having a slight reddish glow. In other words, this object is glowing “red
hot.” Blacksmiths use this fact to determine when a piece of metal has
reached an appropriate temperature, and the necessity of seeing a faint
glow from an object is one reason that blacksmiths often work in dim, low-
light conditions.
For the 6,000-K object, most of the photons emitted fall within the
visible range (Figure 3.2c). Our Sun is, to a good approximation, a 6,000-
K blackbody, and the distribution of photons from the Sun is closely
approximated by this blackbody spectra. Because being able to see confers
a strong advantage in surviving, it is no surprise that the eyes of humans
and other animals have evolved to see this range of wavelengths. In fact,
the human eye is maximally sensitive to light with a wavelength near 0.5
µm, which is the λmax for a 6,000-K blackbody. The chlorophyll molecule,
the key component of photosynthesis, strongly absorbs photons in the
visible range, showing that plants have also evolved to take advantage of
photons emitted by the Sun.
Finally, if the photons emitted by room-temperature objects are not
visible to our eyes, how can we see room-temperature objects, such as this
page? What you see when you look at a room-temperature object are
visible photons (emitted by the Sun or a light bulb or some other hot
object) that have bounced off the object.
An everyday object that uses a lot of the concepts that we have
discussed in this chapter is the humble incandescent light bulb. An
incandescent light bulb consists of a glass envelope containing a small
filament made of a metal, such as tungsten. When the light bulb is turned
on, electricity flows through the filament, heating it to around 3,000 K
(Figure 3.3).
100

Figure 3.3 A schematic of a typical incandescent light bulb.
Figure 3.4 shows the wavelength distribution of photons emitted by a
3,000-K blackbody. As the figure shows, the filament is hot enough that
some of the photons emitted are visible – so humans will see the light bulb
glowing and you can use it to light your room. However, 85 percent of the
photons emitted have wavelengths in the infrared, too long for the human
eye to detect. These infrared photons provide no lighting for humans and
so the energy to produce them is essentially wasted. This makes
incandescent bulbs inefficient as light sources.
Figure 3.4 Emissions spectrum for a 3,000-K blackbody, a typical
filament temperature for an incandescent light bulb. The numbers on the
101

y-axis are omitted.
One way for a light bulb to produce a higher fraction of visible
photons is to run the filament at a higher temperature. As described by
Equation 3.1, this shifts the distribution of emitted photons to shorter
wavelengths, thereby making a greater fraction of them visible to humans
– thereby making the bulb more efficient. The optimal temperature for the
filament would be about the temperature of our Sun, nearly 6,000 K,
which provides the best overlap between blackbody emission and the
human visual range. Unfortunately, at such a temperature, the filament
would immediately vaporize and the bulb would be destroyed.
In fact, conventional incandescent bulbs are run at about as high a
temperature as they can be. To further increase the filament temperature,
the nitrogen and argon found in most bulbs is replaced with halogen gas.
Because of chemical reactions between the halogen gas and the filament,
the filament in these so-called halogen light bulbs can survive at
temperatures several hundred degrees hotter than a regular incandescent
bulb. This means that halogen light bulbs put out more visible photons,
making them more efficient than regular incandescent bulbs.
Unfortunately, because the filament is run so hot, the light bulb itself also
gets extremely hot, creating a fire and burn hazard.
A better way to obtain high efficiency lighting is to change the
technology. Compact fluorescent light bulbs (CFL) and light emitting
diode (LED) light bulbs use different technologies (which I will not
discuss here) to emit most of the bulb’s photons in the visible wavelength
range. The net result is a bulb that is at least five times more efficient. In
other words, a 12-W CFL will produce the same amount of visible light as
a 60-W incandescent light bulb – and it does this by reducing the amount
of infrared light emitted. In an effort to boost energy efficiency, the U.S.
102

Congress passed a law in 2007 phasing out production of standard
incandescent bulbs in the United States by 2014. By the time you read this,
it may be difficult to find those bulbs being sold.
Not only does the wavelength of emission change with temperature,
but the total power emitted also increases with temperature. The observant
reader would have seen this in the varying y-axis ranges in Figure 3.2, but
it is explicitly shown in Figure 3.5a, which shows four different
blackbody-emission curves on a single plot. The plot shows that, at every
wavelength, warmer objects emit more power than cooler objects.
Figure 3.5 Plots of (a) the distribution of power emitted by a blackbody
at four different temperatures (1,600, 1,400, 1,200, and 1,000 K), in
(W/m2)/µm, and (b) total power emitted by a blackbody as a function of
temperature, in W/m2.
For a different view of this, Figure 3.5b plots the total power emitted
by a blackbody as a function of temperature. It is clear that, as the
103

(3.2)
temperature of the object increases, so does the power emitted. It turns out
that there is a simple relation, known as the Stefan-Boltzmann equation,
between the total power radiated by a blackbody and temperature:
P/a is the power emitted by a blackbody per unit of surface area, with
units of watts per square meter; σ is the Stefan-Boltzmann constant, σ =
5.67 × 10−8 (W/m2)/K4; and T is the temperature of the object in Kelvin. If
you multiply P/a by the surface area a of the object (in square meters),
then you get the total power emitted by a blackbody, in watts.
The Stefan-Boltzmann equation has wide applications. By measuring
the amount of power emitted by an object, astronomers use it to infer the
temperature of distant stars and planets. The U.S. military uses the
equation to build sensors to identify and lock onto hot jet engines against a
cold sky in the guidance systems of heat-seeking missiles. And the ear
thermometer that might be in your medicine cabinet right now uses this
same physics to convert the infrared emission from the eardrum and
surrounding tissue into an estimate of body temperature.
As a further example, Figure 3.6 shows an image of my dog, Kasper,
in the infrared. To construct this image, the temperature is determined by
measuring the infrared emission and converting this to temperature. Bright
colors indicate warm temperatures and dark colors indicate cool
temperatures. Like humans, dogs are mammals and their body temperature
is around 38°C. Fur is an insulator, however, so fur-covered regions of the
dog tend to be cooler. Areas that are not fur covered, such as the eyes, are
close to the dog’s internal temperature. Note also the dog’s cold nose.
104

Figure 3.6 Photo of Kasper Dessler in the infrared, with colors assigned
to different temperatures.
An example: How fast is a room-temperature basketball losing
energy by the emission of photons?
At room temperature, a blackbody is emitting σ(300 K)4 = 460
W/m2. A basketball with a radius of 5 in. = 0.13 m has a surface
area of 4π(0.13 m)2 = 0.2 m2. The total rate of energy loss from a
room-temperature basketball as a result of blackbody photon
emission is therefore 460 W/m2 × 0.2 m2, or 92 W. This is about
the same emitted power as a typical light bulb. Of course, you
cannot light a room with a basketball because the photons emitted
by a basketball are outside the range that humans can see.
105

3.4 Energy balance
One of the cornerstones of modern physics is the first law of
thermodynamics, which basically says that energy is conserved. If some
object loses some energy, then some other object must gain that same
amount of energy. Furthermore, because photons are just little packets of
energy, the first law tells us that when an object emits a photon, the
object’s internal energy must decrease. And because temperature is a
measure of internal energy, the emission of a photon therefore causes the
object to cool. Similarly, if a photon hits an object and is absorbed, then
the energy of the photon is transferred to the object’s internal energy and
the object warms.
An example: conservation of money
A good analogy for energy balance is money balance. If you gain
one dollar, then someone else must be one dollar poorer because
money, like energy, cannot be created or destroyed.4
Consider, for example, a checking account. Funds, such as
your paycheck or a birthday check from your grandmother, are
periodically deposited into the account. At the same time, funds are
withdrawn, to pay for things such as rent or a cell phone bill. The
change in your bank balance is equal to the difference between the
total deposits (money in) and total withdrawals (money out). In
equation form, we write this as follows:
106

If money in exceeds money out, that is, your deposits exceed your
withdrawals, then the change in balance is positive and your
balance increases. If money out exceeds money in, then the change
in balance is negative and your balance decreases. If money in and
money out are equal, the change in balance is zero and your
balance is unchanged. This is basically the calculation we do when
we balance our checkbooks.
If the energy flowing into an object (energy in) exceeds the energy
flowing out (energy out), then the internal energy (and temperature) of the
object increases. Written mathematically, this is as follows:
Here the symbol α means “is proportional to.” Note that, if energy in and
energy out are equal, the internal energy and temperature are unchanged.
We call this situation equilibrium.
A good example that draws many of the concepts in this chapter
together is your home oven. Most people, if asked how an oven cooks,
would answer, “Because it’s hot inside.” However, you may be surprised
that the physics is subtler than you realize. Ovens do not cook because the
air in the oven is hot – air is a terrible conductor of heat. Rather, ovens
cook by infrared radiation.
When an electric oven is turned on, electricity runs through a heating
element. The element heats up, eventually reaching temperatures high
enough that it glows a dark orange. At this point, the element is radiating
an enormous amount of power, typically several thousand watts.
The photons emitted by the heating element are absorbed by the walls
of the oven, heating them. When the walls reach a predetermined
temperature, typically 350–450°F (450–500 K), then the oven is
107

“preheated” and the cook puts the food, say a turkey, into the oven. Let us
assume the turkey came out of the refrigerator and has a temperature of
3°C or 276 K. At this temperature, the turkey is radiating 330 W/m2. If the
turkey has a surface area of 0.1 m2, then the total power radiated by the
turkey is 33 W.
The turkey is also absorbing photons from the oven’s hot walls. The
oven walls, at 375°F (465 K), are radiating 2,650 W/m2. The total surface
area of the oven’s six walls is approximately 1.3 m2, so the total power
radiated by the oven’s wall is roughly 3,500 W. Most of the energy
radiated by the oven’s walls misses the turkey in the middle and hits the
other walls, and only a fraction of photons emitted by the walls hits the
turkey. The turkey absorbs photons emitted by an area of the walls equal to
the surface area of the turkey, 0.1 m2. Given that the walls emit 2,650
W/m2, that means that the turkey is absorbing 265 W of power.
Because the turkey is emitting 33 W but absorbing 265 W, the
internal energy of the turkey is increasing and it is therefore warming.
Eventually, the turkey reaches the temperature when it is considered
“done,” and the cook removes it from the oven. That is how a conventional
oven cooks.
While the turkey is absorbing energy from the walls, by conservation
of energy the walls must be losing energy and cooling down. The oven has
a thermostat in it that senses this cooling and turns on the heating element
to maintain the wall temperature at 375°F. This occasional cycling back on
of the heating element is familiar to any cook.
A microwave oven also cooks food by bombarding food with
photons. However, instead of bombarding the food with infrared photons,
a microwave oven bombards the food with microwave photons, which
have longer wavelengths. For reasons that we will not go into here,
microwave ovens cook faster because they are able to deliver higher rates
108

of power to the food than a conventional oven can. In the example here,
the oven is delivering 265 W of power to the turkey. By using
microwaves, however, the oven is able to deliver five to ten times that
amount. The net result is that the food is heated more rapidly in a
microwave oven than it is in a conventional oven.
I hope that you have a sense of the importance of the physics we have
discussed in this chapter – it has a profound impact on your life and the
world around you. I will show you in the next chapter that it also plays a
key role in climate.
109

3.5 Chapter summary
Energy is expressed in units of joules (J). Power is the rate that
energy is flowing, and it is expressed in watts (W); 1 W = 1 J/s.
Temperature is a measure of the internal energy of an object and is
frequently expressed by physicists in units of Kelvin. The
temperature in Kelvin is equal to the temperature in degrees
Celsius plus 273.15.
Photons are small discrete packets of energy. They have a
characteristic size, known as the wavelength, which determines
how the photons interact with matter. Photons with wavelengths
between 0.3 and 0.8 m are visible to humans; photons with
wavelengths between 0.8 and 1,000 µm are called infrared.
Most objects emit blackbody radiation. The characteristic
wavelength emitted by a blackbody is equal to 2,897⁄T (where
wavelength is in microns and temperature is in Kelvin). The total
power emitted per unit area by a blackbody is equal to σT4, where
σ = 5.67 × 10−8 (W/m2)/K4 and temperature is in Kelvin. Photons
emitted by room-temperature objects are in the infrared and are not
visible to humans.
When a photon is emitted by an object and then absorbed by
another object, this process transfers a small amount of energy
from the emitter to the absorber.
If the energy received by an object by absorbing photons exceeds
the energy lost by emitting photons, then the object’s internal
energy increases – and it warms up. The object cools off if the
110

energy in emitted photons exceeds the energy received by
absorbing photons.
111

Additional reading
D. Archer, Global Warming: Understanding the Forecast, 2nd ed.
(Hoboken, NJ: Wiley, 2011). Chapter 2 of this excellent introductory text
on climate change covers blackbody radiation.
For a more detailed discussion of blackbody radiation, see almost any
introductory physics book.
See www.andrewdessler.com/chapter3 for additional resources for the
chapter and www.andrewdessler.com/computer for computer exercises
that illustrate some of the important concepts of radiation and energy
balance.
112

http://www.andrewdessler.com/chapter3

http://www.andrewdessler.com/computer

Terms
Blackbody
Electromagnetic radiation
Emissions spectra
Energy
Energy balance
Equilibrium
Incandescent light bulb
Infrared radiation
Internal energy
Joule
Kelvin scale
Micron
Photons
Power
Temperature
Ultraviolet
Visible photons
Watt
Wavelength
113

Wien’s displacement law
114

Problems
1. The temperature of an object goes up by 1 K. How much did it go
up in degrees Fahrenheit and how much in degrees Celsius?
2. A sphere with a radius of 1 m has a temperature of 100°C. How
much power is it radiating?
3. As a room-temperature object increases in temperature, it begins to
glow. Describe the progression in colors as the object heats up.
Ultimately, what happens to the glow if the warming continues to
nearly infinite temperatures?
4. Consider two stars that have the spectra shown in Figure 3.7. Based
just on the information provided in this plot, what are the colors and
radiating temperatures of the stars? (The gray shading shows the
range of wavelengths that humans can see.)
5. How much total energy (in watts) is the Sun radiating? It is a
6,000-K blackbody with a radius of 700,000 km.
6. You can dim an incandescent bulb by decreasing the temperature
of the filament. What do you think happens to the color of the bulb as
it dims? Find a dimmer and test your hypothesis.
7. If you run a 60-W light bulb for one week, how many joules of
energy have been consumed?
8. Why are incandescent light bulbs being phased out in many
countries (including the United States)?
9. The Sun as a blackbody:
115

a. The Sun is a 6,000-K blackbody. At what characteristic
wavelength does it radiate?
b. How much power per unit surface area is the Sun radiating?
c. Imagine that the Sun had a radius twice as large as it presently
does, but emitted the same total amount of energy. What
temperature would the Sun have to be?
10. Ein is the energy being absorbed by an object, and Eoutis the
energy being radiated.
a. If the temperature of an object is not changing, what does this
tell us about Ein and Eout?
b. If the temperature of an object is increasing, what does this tell
us about Ein and Eout?
11. Your bank account has the same balance on April 1 as it did on
March 1. Your friend suggests that this means that you did not
deposit or withdraw any money for the entire month. Is that correct?
Explain why or why not.
12. Heat capacity is the amount of energy it takes to warm up an
object by 1 K. The heat capacity of water is 4.18 J/g/K; in other
words, if you add 4.18 J to 1 g of water, the water will warm by 1 K.
Imagine you have a cup containing 200 g of water that is absorbing
150 W of power.
a. At what rate is the water warming? Answer is degrees K per
second.
b. If the cup starts at room temperature, how long would you have
to heat it to reach boiling?
To verify your answers, make sure the units work out.
116

13. I mentioned in the chapter how microwave ovens are able to
deliver more energy to food during cooking than conventional ovens,
so microwave ovens can cook food faster. For the following, imagine
you are cooking a turkey in a conventional oven at 325°F.
a. What would you have to do in order to increase the amount of
power being delivered to the turkey with the conventional oven?
b. Would this cook the turkey faster? Why do we not cook turkeys
that way?
c. Why are microwave ovens able to deliver so much energy to
food, while conventional ovens cannot?
(to answer parts b and c, you have to know that the energy from
a photon is absorbed by an object over a layer about one
wavelength thick)
Figure 3.7 Emissions spectra of two hypothetical stars.
1 Note that you do not use a degree sign (°) when writing temperature in
Kelvin. Thus, room temperature is 295 K, not 295°K. That is because
Kelvin is an absolute temperature scale and 0 K is absolute zero.
2 When people hear the word radiation, they often think of nuclear
radiation. Such radiation has very high energies because it originates
117

from changes in atomic nuclei, and as a result this radiation can cause
cancer and other medical problems. Electromagnetic radiation discussed
here generally originates from changes in the atoms’ electrons or from
changes in the molecule’s rotational or vibrational state and therefore
has far less energy – so it is generally not a health hazard. This is good,
because you are surrounded by electromagnetic radiation right now.
3 Electromagnetic radiation also behaves like a wave, but for this
problem it is easier to think of it as a particle.
4 This rule does not apply to national governments, which can print
money.
118

4
A simple climate model
◈
Scientists have been studying the Earth’s climate for nearly 200 years and,
over that time, a sophisticated and well-validated theory of our climate has
emerged. In this chapter, we take the fundamental physics we learned in
the last chapter and use it to explain how greenhouse gases warm the
planet and why the temperature of the Earth is what it is. By the end of the
chapter, you will understand why scientists have such high confidence that
adding greenhouse gases to the atmosphere will warm the planet.
119

4.1 The source of energy for our climate
system
The first step to understanding the climate is to do an energy budget
calculation, which requires us to calculate the energy in and energy out for
the Earth. The ultimate source of energy for our planet is the Sun, which
puts out an amazing 3.8 × 1026 W (380 trillion trillion W) of power. The
Sun emits photons in all directions, so only a small fraction of this energy
falls on the Earth. So the first step in calculating energy in is to determine
the intensity of sunlight at the Earth’s orbit.
To estimate this, imagine a sphere surrounding the Sun, with a radius
equal to the Sun-Earth distance, 150 million km (Figure 4.1). Because the
sphere completely encloses the Sun, all of the sunlight emitted by the Sun
must fall on the interior of the sphere. The surface area of the sphere is
4πr2 = 4π(150 million km)2 = 2.8 × 1017 km2 = 2.8 × 1023 m2. Dividing
the total energy emitted by the Sun by the area of the sphere produces an
estimate of the intensity of solar radiation at the Earth’s orbit: 3.8 × 1026
W/2.8 × 1023 m2 = 1,360 W/m2. This value, 1,360 W/m2, is known as the
solar constant for the Earth; it is frequently represented in equations by the
symbol S.
120

Figure 4.1 Solar constant calculation: A sphere (gray) surrounds the
Sun with a radius equal to the Earth’s orbit (dashed line); all radiation
emitted by the Sun (black arrows) falls on this sphere.
As should be obvious, the solar constant is a function of how far the
planet is from the Sun. As a planet gets closer to the Sun, the solar
constant for that planet increases; if it gets further away, the solar constant
decreases. Our next-door neighbor Venus is located 107 million km from
the Sun. Thus, the solar constant for Venus is 3.8 × 1026 W divided by the
surface area of a sphere with a radius of 107 million km, 1.43 × 1023 m2 –
which yields a value of 2,600 W/m2.
Now that we know the Earth’s solar constant, we can determine the
total solar energy falling on the Earth. The easiest way to quantitatively
calculate this is to realize that, if we set up a screen behind the Earth, the
Earth would cast a circular shadow on the screen, with a radius equal to
the radius of the Earth (Figure 4.2). The amount of sunlight falling on the
Earth is equal to the amount that would have fallen into the shadow area if
the Earth were not there. The shadow area is πR2, where R is the radius of
121

the Earth. So the total solar energy that would have fallen into that area is
πR2 times the solar constant S.
Figure 4.2 The Earth is casting a shadow on a screen placed right
behind it because it blocks sunlight. The total amount of solar energy
falling on the Earth is the same as what would have fallen into the
shadow area.
Given that the radius of the Earth is approximately 6,400 km = 6.4 ×
106 m, and S = 1,360 W/m2, solar energy is falling on the Earth at a rate of
1.8 × 1017 W or 180,000 TW (1 TW, called a terawatt, is 1012 or a trillion
watts). This is an immense amount of power. Human society today
consumes about 16 TW, so this simple calculation shows why solar energy
is the Holy Grail of renewable energy: If we could capture just 0.01
percent of the solar energy falling on the Earth, we could satisfy all of the
world’s current energy needs.
Not all of the photons from the Sun that fall on the Earth are absorbed
by it. Some of the photons are reflected back to space by clouds, ice, and
other reflective elements of the Earth system. The reflectivity of a planet is
called the albedo, from the Latin word for “whiteness” (the word albino
derives from the same root). It is frequently represented by the symbol α,
which is the fraction of incident photons that are reflected back to space;
for the Earth, α is 0.3. This means that 1 − α is the fraction of photons that
122

(4.1)
(4.2)
are absorbed by the Earth. Taking this into account, energy in (Ein) for the
Earth is
Evaluating Equation 4.1 yields an estimate of Ein for the Earth of 120,000
TW. In the rest of the chapter, we will find it more useful to do the
calculation per square meter of the Earth’s surface area, so we divide
Equation 4.1 by the surface area of the Earth, 4πR2:
Note that the πR2 terms cancel, so the net amount of solar energy absorbed
per square meter is not a function of the Earth’s size. Plugging values of S
= 1360 W/m2 and α = 0.3 into Equation 4.2, we obtain a value of 238
W/m2 for the Earth’s Ein. This is a good number to remember.
You might have noticed that I have become a bit sloppy with the
terms energy and power in the previous discussion. In Equation 4.1, for
example, the mathematical abbreviation for energy in appears on the left-
hand side, but the right-hand side has units of power (Watts). The physics
pedants will argue that we should be writing power in instead of energy in,
and they are indeed correct. However, my choice of terminology here
reflects the terminology actually used by scientists who do these kinds of
energy-balance calculations. If you go to a meeting of climate scientists or
read the peer-reviewed climate literature, you will find that they use power
and energy interchangeably in equations like Equation 4.2. If you worry
about things like this, then just keep track of the units and you will always
know what is being talked about.
So the Earth absorbs an average of 238 W/m2 from the Sun, but that
does not mean that every square meter of the Earth absorbs this amount. In
123

fact, the amount of solar energy absorbed varies widely across the planet.
First, the nighttime half of the Earth receives no energy from the Sun at all.
Second, the amount falling on a square meter of the daytime half is
determined by the orientation of that square meter with respect to the
incoming beams of sunlight. The amount of sunlight received is at
maximum if the surface is oriented perpendicular to the incoming beam
(Figure 4.3a). As the surface rotates away from perpendicular, the amount
of solar energy intercepting the surface decreases (Figure 4.3b), eventually
reaching zero for a surface parallel to the incoming beam (Figure 4.3c).1
Figure 4.3 Schematic showing how the amount of energy falling on a
surface is dependent on the angle between the surface and the incoming
beams of light: (a) perpendicular, (b) rotated away from perpendicular,
and (c) parallel.
Figure 4.4 shows how this leads to variations in the amount of solar
energy falling on the Earth’s surface with latitude. In the tropics (Arrow
A), the surface of the Earth is perpendicular to the incoming solar light
beams, corresponding to the situation in Figure 4.3a. The surface in the
mid-latitudes (Arrow B) is at a moderate angle to the incoming solar light
beams, corresponding to the situation in Figure 4.3b. This means that mid-
latitudes receive less solar radiation per square meter than the tropics.
124

Finally, the polar regions (Arrow C) correspond to the situation in Figure
4.3c, so this region receives even less solar energy.
Figure 4.4 Schematic showing how the amount of solar energy falling
on a square meter of the Earth’s surface is determined by the latitude.
In addition to variations in the incoming sunlight with latitude, the
albedo of the planet also varies widely. The tropics are mainly open ocean,
which is dark and therefore has a low albedo. Combined with the large
amount of solar energy per square meter, the tropics therefore experience
far more solar heating than anywhere else on the planet. The high latitudes,
however, are generally covered by snow and ice, giving them a high
albedo. Combined with the small amount of solar energy received per
square meter, this means that the polar regions absorb the least amount of
solar energy. This provides us with a simple but fundamentally correct
explanation of why the tropics tend to be the warmest place on the planet
and the polar regions the coldest.
125

(4.3a)
(4.3b)
4.2 Energy loss to space
In the early nineteenth century, Joseph Fourier, one of history’s great
mathematicians, asked a deceptively simple question: Because energy is
always falling on the Earth from the Sun, why does the Earth not heat up
until it is the same temperature as the Sun? The answer he determined is
that the Earth is losing energy at a rate equal to the rate at which it is
receiving energy from the Sun.
On the basis of what we learned in Chapter 3, you may rightly guess
that the Earth loses energy back to space by means of blackbody radiation
(Figure 4.5). For a blackbody, P/a is σT4, where P/a is the power emitted
per square meter, T is the temperature of the planet, and σ is the Stefan-
Boltzmann constant, 5.67 × 10–8 (W/m2)/K4. Setting Ein (Equation 4.2)
equal to P/a, the rate of energy out, we get the following equation:
Solving for T, we get
Plugging S = 1,360 W/m2 and α = 0.3 into Equation 4.3b yields2 a
temperature T = 255 K (−18°C). The actual average temperature of the
Earth is closer to 288 K (15°C), so our estimate of the Earth’s temperature
is too cold by 33°C. Where did our calculation go wrong? It turns out that
what we have neglected is the heating of the planet by the Earth’s
atmosphere, which is frequently referred to as the greenhouse effect.
126

Figure 4.5 Diagram of the Earth, showing the emission of infrared
blackbody radiation in all directions.
127

4.3 The greenhouse effect
128

4.3.1 One-layer model
To understand the impact of the atmosphere on our planet’s temperature,
let us make the following assumptions (which turn out to be reasonably
accurate).
1. The Earth’s atmosphere is transparent to visible photons emitted by
the Sun (which have wavelengths from 0.3–0.8 µm), so these photons
pass through the atmosphere and are absorbed by the surface.
2. The atmosphere is opaque to infrared photons emitted by the
surface (wavelengths longer than 4 µm), and so all of these photons
are absorbed by the atmosphere.
3. The atmosphere also behaves like a blackbody, so it emits photons
based on its temperature. It emits photons equally both upward and
downward.
4. Photons emitted by the atmosphere in the upward direction escape
to space and carry energy away from the Earth. Photons emitted
downward are absorbed by the surface.
This one-layer model is diagramed in Figure 4.6. For conceptual
simplicity, the diagram shows the effects of the atmosphere concentrated
in a single thin layer, which is why this model is frequently called a “one-
layer” model.
129

Figure 4.6 Schematic of energy flow on a planet with a one-layer
atmosphere. The atmosphere is represented by a single layer that is
transparent to visible photons but absorbs all infrared photons that fall
on it. The arrows show global average energy flows with values in
W/m2.
To calculate the surface temperature in this model, we assume that the
planet as a whole, as well as the surface and the atmosphere individually,
must all be in energy balance (where Ein equals Eout). First, let us consider
the energy balance for the planet as a whole. Energy in to the planet is
coming entirely from the Sun. Energy out to space is coming entirely from
the atmosphere (remember: any photons emitted by the surface are
absorbed by the atmosphere; they do not escape to space). Using the
Earth’s values of solar constant and albedo, the energy in from the Sun is
238 W/m2 (Equation 4.2). This means that the atmosphere must be
radiating 238 W/m2 upward to space in order for the planet as a whole to
be in energy balance. Because the atmosphere radiates equally upward and
downward, the atmosphere is also radiating 238 W/m2 back toward the
Earth’s surface.
Now let us consider energy balance for the surface. Energy in for the
surface is 238 W/m2 from the Sun and 238 W/m2 from the atmosphere, for
a total of 476 W/m2. This means that the surface has to be emitting 476
W/m2 upward in order to achieve energy balance.
To make sure we did not make a mistake, we can check the energy
balance for the atmosphere. Energy in comes from the surface, which is
emitting 476 W/m2. Energy out comes from emission of 238 W/m2
upward to space and 238 W/m2 downward to the surface, for a total energy
out of 476 W/m2. Thus, the atmosphere is indeed in energy balance.
130

So what is the temperature of the surface? If we know that the surface
is emitting 476 W/m2, we can determine its temperature by using the
Stefan-Boltzmann equation from Chapter 3: Eout = P/a = σT4. Solving σT4
= 476W/m2 for T yields a surface temperature of 303 K (30°C), which is
48°C warmer than that for the planet without an atmosphere.
This is an incredibly important result: the addition of an atmosphere
that is opaque to infrared radiation has significantly warmed the planet’s
surface. Conceptually, this occurs because the surface of the planet with an
atmosphere is heated not just by the Sun but also by the atmosphere. Of
course, if you walk outside, you cannot see the atmosphere heating the
Earth’s surface because the photons it emits are not visible, but they still
carry energy. When scientists talk about the greenhouse effect, it is this
heating of the surface by the atmosphere to which they are referring.
An alternative way to think about the greenhouse effect is that the
atmosphere warms the surface by making it harder for the surface to lose
energy to space. Without an atmosphere, all of the photons emitted by the
surface escape to space; the surface has to emit only 238 W/m2 for the
planet to be in energy balance. With a one-layer atmosphere, though, only
half of the photons emitted by the surface end up escaping to space – the
other half are returned to the surface. This means that the surface must
emit twice as much, 476 W/m2, in order for 238 W/m2 to escape to space.
This higher rate of emission requires a warmer surface.
131

4.3.2 Two-layer model
Now let us consider a planet with a “two-layer” atmosphere (Figure 4.7).
Once again, the atmosphere is transparent to visible radiation but opaque
to infrared. That means that photons from the Sun pass through the
atmosphere and are absorbed by the surface. Photons emitted by the
surface are absorbed in the lower atmosphere. Photons emitted by the
lower atmosphere in the upward direction are absorbed by the upper
atmosphere; photons emitted in the downward direction are absorbed by
the surface. Photons emitted by the upper atmosphere in the upward
direction escape to space; photons emitted in the downward direction are
absorbed by the lower atmosphere.
Figure 4.7 Schematic of energy flow on a planet with a two-layer
atmosphere, with values in W/m2.
Once again, the key to determining the surface temperature is to
enforce energy balance for the planet as a whole, the surface, and both
atmospheric layers. The easiest way to do this is to start with planetary
energy balance and then work downward from the topmost layer to the
surface. Planetary energy balance requires energy out for the planet to
132

balance energy in from the Sun. Because energy out comes entirely from
the upper layer, it must be emitting 238 W/m2 to space in order to balance
the 238 W/m2 that the Sun is providing the planet (assuming terrestrial
values for S and α). That, in turn, means that the upper layer is also
emitting 238 W/m2 downward.
Totaling the emissions in both directions, the upper layer is emitting
476 W/m2. Because energy out must equal energy in for the layer, this
layer must be receiving 476 W/m2 from the lower atmospheric layer. Thus,
we know that the lower layer is emitting 476 W/m2 upward – and therefore
downward, too. For the lower layer to achieve energy balance, the lower
layer must also be receiving 476+476 = 952 W/m2. We already calculated
that 238 W/m2 are coming from the upper layer, so that means that 714
W/m2 must be coming from the surface to the lower layer.
We can verify our result by examining the energy balance for the
surface. The surface receives 476 W/m2 from the lower atmosphere and
238 W/m2 from the Sun, for a total Ein of 714 W/m2. This corroborates
what we calculated must be Eout for the surface based on energy balance
for the lower atmosphere.
Finally, for the surface to be emitting 714 W/m2, its temperature must
be 335 K (62°C). This is 32°C warmer than the surface of the planet with a
one-layer atmosphere and 80°C warmer than a planet with no atmosphere.
Thus, adding a second layer to the atmosphere further increases the
planet’s surface temperature.
133

4.3.3 n-layer model
Now let us derive the surface temperature for a planet with n layers
(Figure 4.8). For some variety, let us assume that the planet has a solar
constant S = 2,000 W/m2 and an albedo α = 0.7. Thus, energy in for this
planet is S(1 − α)/4 = 150 W/m2. This means that upward emissions from
the topmost layer of the atmosphere (Layer 1) must also be 150 W/m2.
And because upward and downward emissions must be the same, this layer
is also emitting 150 W/m2 downward, so that total energy out for this layer
is 300 W/m2. This in turn means that energy in for Layer 1 must also be
300 W/m2.
Figure 4.8 Schematic of energy flow on a planet with an n-layer
atmosphere; layers are numbered from 1 to n (topmost to bottommost
layers), with values in W/m2.
Energy in for Layer 1 comes entirely from energy emitted by Layer 2.
Layer 2 must therefore be emitting 300 W/m2 upward. This means that it is
also emitting 300 W/m2 downward, for a total energy out of 600 W/m2.
Energy in for Layer 2 comes from downward emissions of Layer 1 and
upward emissions of Layer 3 and must total 600 W/m2 in order to balance
134

(4.4)
energy out. Downward emissions from Layer 1 are 150 W/m2, which
means that upward emissions from Layer 3 must be 450 W/m2.
Layer 3 must be emitting 450 W/m2 both upward and downward, for
a total energy out of 900 W/m2. Energy in from downward emissions from
Layer 2 is 300 W/m2, meaning that upward emissions from Layer 4 must
be 600 W/m2.
By this time, a pattern has emerged and we can extrapolate to the
bottommost layer, layer n. Layer n is emitting 150n in both upward and
downward directions. This in turn means that the surface is receiving 150n
emitted from the bottommost layer and 150 W/m2 from the Sun. For
energy to balance, the surface must be emitting 150(n + 1) W/m2 upward.
Setting 150(n + 1) = σT4, we can solve for the surface temperature T of
this planet:
Figure 4.9 shows the surface temperature as a function of n, calculated by
using Equation 4.4. As you increase the number of layers, the surface gets
hotter and hotter. However, the warming is not linear – each additional
layer produces less warming than the previous layer.
Figure 4.9 Surface temperature for the n-layer planet, as a function of
the number of layers. Here, S W/m2.
135

(4.5)
We can also write the general solution for the surface temperature of
an n-layer planet:
This is an important equation and one that is good to memorize. It says
that the surface temperature of the planet is basically determined by three
parameters: the number of layers in the atmosphere (n), the solar constant
(S), and the albedo (α). To connect this equation to the real world, I should
make clear what the “number of layers” physically represents. As I will
discuss in Chapter 5, it is the greenhouse gases in our atmosphere that
absorb infrared photons. And the number of layers is equivalent to the
amount of greenhouse gas in the atmosphere. Therefore, an increase in the
amount of greenhouse gas in the atmosphere corresponds to an increase in
the number of layers – and a warming climate.
Before we proceed to testing our theory, it is important to reiterate
two important points:
1. We can now answer the question, “Why is the Earth’s temperature
what it is?” The temperature of our climate system is set by the
requirement that energy in and energy out balance. And if n, S, or α
change, then the temperature will adjust as required to reestablish this
energy balance.
2. As you add more greenhouse gases to the planet (i.e., increase n),
the temperature of the planet will increase.
136

4.4 Testing our theory with other
planets
It is important to emphasize that the n-layer model discussed in Subsection
4.3.3 makes many simplifying assumptions. For example, not all energy
transport on a planet is by radiation – some is transported by atmospheric
motions, such as thunderstorms. Another simplification is that the model
assumes an infinitely fast horizontal energy transport, allowing us to use a
single temperature for the planet. In reality, though, transport of energy is
slow enough that large temperature differences can develop between
regions (e.g., the tropics and the polar regions or between night and day).
This means that you should not expect the model to produce
quantitatively accurate surface temperatures. Nonetheless, the model
captures the essential physics of our climate system. And, as I will show in
this section, the model is successful in making qualitative predictions of
the relative surface temperatures of the Earth and its nearby neighbors,
Mercury, Venus, and Mars.
Table 4.1 lists the important characteristics of the planets, and it
reveals some puzzles. Mercury is the planet closest to the Sun, yet Venus,
twice as far from the Sun as Mercury, has a surface temperature that is
approximately 300 K warmer. This result becomes even more puzzling
when we realize that, because of its high albedo, the energy in for Venus,
S(1 − α)⁄4 = 200 W/m2, is more than a factor of ten smaller than that for
Mercury (2,250 W/m2). It is even less than the energy in for the Earth (238
W/m2) – yet Venus is 450 K hotter than the Earth.
Table 4.1 Data on the four inner planets in our solar system
137

Planet Solar
constant
(W/m2)
Albedo Observed
surface
temperature
(K)
Inferred n
Mercury 10,000 0.1 452 0.052
Venus 2,650 0.7 735 82
Earth 1,360 0.3 289 0.65
Mars 580 0.15 227 0.22
Given the surface temperature, albedo, and solar constant, we can
solve Equation 4.5 for n, which is the number of layers required to satisfy
energy balance. This “inferred n” is also listed in Table 4.1. The inferred n
for Mercury is near zero, suggesting it has almost no greenhouse effect.
This is correct, because Mercury has essentially no atmosphere. For Mars,
inferred n = 0.22. This again makes some sense – Mars has a thin
atmosphere mostly containing carbon dioxide, so it does have some
greenhouse effect. However, the Martian atmosphere has fewer
greenhouse gases than the Earth’s atmosphere, so the greenhouse effect on
Mars is expected to be weaker than that on Earth. Our calculations confirm
that.
Finally, our calculations suggest that Venus, with inferred n = 82, has
a massive greenhouse effect. This is again correct. The surface pressure on
Venus is ninety times that of Earth (1,300 psi, or pounds per square inch,
compared with 14.5 psi here on Earth), and the atmosphere is mainly
composed of carbon dioxide. The result of this massive, greenhouse-gas-
rich atmosphere is a planet hotter than the inside of your oven on broil –
hot enough even to melt lead. Thus, we see that Equation 4.5 successfully
explains the relative climates of the innermost planets of our solar system.
138

4.5 Chapter summary
In this chapter, we created a very simple climate model based on
the fact that the solar energy received by a planet (Ein) must be
balanced by the energy that is radiating to space (Eout). The
temperature of the planet adjusts until this balance is achieved.
For a planet, Ein = S(1 − α)/4. S is the solar constant, which is the
intensity of sunlight at the planet’s orbit (in units of W/m2), and α is
the planet’s albedo, which is the fraction of photons that fall on the
planet that are reflected back to space.
The energy out for a planet is due to blackbody radiation; Eout =
σT4.
In our simple model of the climate, the atmosphere is entirely
transparent to visible radiation from the Sun, but it absorbs all
infrared radiation. We can then calculate the surface temperature
by enforcing energy balance for the surface, the atmosphere, and
the planet as a whole.
We derived a general equation, Equation 4.5, for the surface
temperature T of a planet. It is repeated here:
This equation says that the surface temperature of the planet is
determined by three parameters: the number of layers in the
atmosphere (n), which is a proxy for how much greenhouse gas is
in the atmosphere, the solar constant (S), and the albedo (α).
139

This simple model also explains the relative temperatures of the
Earth’s nearest neighbors, namely Mercury, Venus, and Mars.
140

Additional reading
For a more complete and technical description of the physics of climate,
see the following books.
D. Archer, Global Warming: Understanding the Forecast, 2nd ed.
(Hoboken, NJ: Wiley, 2011). Chapter 3 provides another description of the
layer model at about the level of this textbook. Check it out to see how
another author explains it.
J. T. Houghton, The Physics of Atmospheres (Cambridge: Cambridge
University Press, 2001). This book covers climate physics at a level
appropriate for an upper-level physics undergraduate.
R. T. Pierrehumbert, Principles of Planetary Climate (Cambridge:
Cambridge University Press, 2011). This book is written at a level
appropriate for a physics graduate student.
See www.andrewdessler.com/chapter4 for additional resources for the
chapter and www.andrewdessler.com/computer for computer exercises
that illustrate some of the important concepts of energy balance and layer
models.
141

http://www.andrewdessler.com/chapter4

http://www.andrewdessler.com/computer

Terms
Albedo
Greenhouse effect
Greenhouse gases
Solar constant
142

Problems
1. What is the surface area of a sphere with radius r? What is the area
of a disk with radius r? What is the area of a disk with diameter d?
2. A planet in another solar system has a solar constant S = 2,000
W/m2, and the distance between the planet and the star is 100 million
km.
a) What is the total power output of the star? (Give your answer in
watts.)
b) What is the solar constant of a planet located 75 million km
from the same star? (Give your answer in watts per square meter.)
3. Draw a diagram (like Figure 4.6) that shows the energy flows for a
planet with a one-layer atmosphere. The solar constant for the planet
is S = 900 W/m2, and the albedo of the planet is α = 0.25. Make sure
each arrow is labeled with the energy flow. What is the surface
temperature of this planet?
4. Draw a diagram (like Figure 4.7) that shows the energy flows for a
planet with a two-layer atmosphere. The solar constant for the planet
is S = 3,000 W/m2 and the albedo of the planet is α = 0.1. Make sure
each arrow is labeled with the energy flow. What is the surface
temperature of this planet?
5. Two people argue about why Venus is so much warmer than the
Earth. The first argues that it is because Venus is closer to the Sun, so
it absorbs more solar energy. The second argues that it is because
143

Venus has a thick, greenhouse-gas-rich atmosphere. Which person is
right, and why is the other one wrong?
6. Some recently discovered planets in other solar systems are so hot
that they glow in the visible; they are literally “red hot” (e.g., do a
Google search for “HD 149026b”).
a) How many atmospheric layers would the Earth need before it
glowed in the visible? (Assume S = 1,360 W/m2 and α = 0.3.) To
answer this, you must first estimate what temperature the Earth has
to be to begin glowing.
b) Alternatively, what would the solar constant have to increase to
for a one-layer planet with an albedo α = 0.3?
c) How far would the Earth have to be from the Sun in order to
have this solar constant?
7. Assume a planet with a one-layer atmosphere has a solar constant S
= 2,000 W/m2 and an albedo α = 0.4.
a) What is the planet’s surface temperature? Make the standard
assumption that the atmosphere is transparent to visible photons
but opaque to infrared photons.
b) During a war on this planet, a large number of nuclear weapons
are exploded, which kicks enormous amounts of dust and smoke
into the atmosphere. The net result is that the atmosphere now
absorbs visible radiation – so solar energy is now absorbed in the
atmosphere. It also still absorbs infrared radiation. Draw a diagram
like Figure 4.6 to show the fluxes for this new situation, and
calculate the planet’s surface temperature. The solar constant and
albedo remain unchanged.
144

c) Explain in words why the temperature changes the way it does
after the nuclear war. Is describing this as “nuclear winter”
appropriate?
8. Assume a planet with a one-layer atmosphere and values of solar
constant S = 1,000 W/m2 and albedo α = 0.25. Let us assume there is
some dust in the atmosphere, so that 50 percent of the Sun’s energy is
absorbed by the atmosphere and 50 percent by the surface. Draw a
diagram like Figure 4.6 to show the fluxes, and calculate the planet’s
surface temperature.
9. Derive an expression for the fraction of energy received by the
surface that comes from the atmosphere (this is the amount of energy
that comes from the atmosphere divided by the sum of energy from
the Sun and energy from the atmosphere). Using values in Table 4.1,
calculate the fraction for Mercury, Earth, and Venus. Make the
standard assumption that the atmosphere is transparent to visible
photons but opaque to infrared photons.
10. On Mercury, which has no atmosphere, the difference in
temperature between daytime and nighttime temperatures can be 700
K. On the Earth, the difference between daytime and nighttime
temperatures can be 30 K. On Venus, there is basically no difference
between daytime and nighttime temperatures. Why is this? (If you get
stuck, working Question 9 might help you answer this question.)
11. As we will discover in Chapter 11, one way to solve global
warming is to increase the reflectivity of the planet (I will explain
how later). To reduce the Earth’s temperature by 1 K, how much
would we have to change the albedo? (assume a one-layer planet with
an initial albedo of 0.3 and solar constant of 1360 W/m2).
145

12. Given fixed n and α, how does the temperature of a planet vary
with r, the distance between the planet and the star? Hint: Work out
how S varies with r, and plug that into Equation 4.5.
13. A planet has a solar constant S = 2,000 W/m2, an albedo α = 0.7,
and a radius r = 3,000 km. What would happen to the temperature if
the planet’s radius doubles?
14. One argument you hear against mainstream climate science is that
adding greenhouse gases to the atmosphere is like painting a window.
Eventually, the window is opaque, so that adding another coat of
paint does nothing. Is this a good analogy? Is there a point where
adding greenhouse gases to the atmosphere does not lead to increases
in the planet’s temperature?
15. Imagine that the Sun’s radius doubles (but the Sun maintains the
same surface temperature). What happens to the Earth’s solar constant
and surface temperature?
16. Explain how the variation of solar energy with surface orientation
(Figure 4.3) can explain the variation in local temperature through the
day. The warmest temperatures during the day are usually found from
3 PM to 5 PM; does this fit with your theory?
17. If you were on a spacecraft and you pointed an infrared
thermometer at a one-layer planet, what temperature would it read?
What if the planet had two layers? Or n-layers? Assume S and α are
the same as for the Earth. Remember that an infrared thermometer
measures the radiation coming from an object and yields the
temperature the object must be in order to be emitting that radiation.
18. Newly formed stars are often obscured by the dense dust clouds
from which they form. To see how these appear to observers on the
146

Earth, imagine that fifty stars, each identical to our Sun, form in a
spherical cloud of dust with a radius of 100 billion km. Much like the
atmosphere, the dust absorbs all of the light given off by the stars and
radiates an equal amount of energy to the rest of the Universe. What
temperature does the cloud appear to be from outside the cloud? What
wavelength telescope would you need to see the dust cloud?
1 For those with a good grasp of geometry and geography, the energy
falling on a square meter of surface varies as the cosine of the latitude.
2 The mathematical equation a = means that a4 = y. To calculate the
fourth root of y on a calculator, you can use the yx key found on most
calculators, where x = 0.25. A simpler way to calculate the fourth root
of y is to take the square root of y and then take the square root of that
number – in other words,
147

5
The carbon cycle
◈
In the simple model of the climate presented in Chapter 4, the temperature
of a planet is set by the number of atmospheric “layers,” the albedo, and
the solar constant. I said there that the number of layers is determined by
the abundance of greenhouse gases in the atmosphere, but I was
intentionally vague about what a greenhouse gas is, or which components
of our atmosphere are greenhouse gases. In this chapter, I address these
questions and discuss in detail one of our atmosphere’s most important
greenhouse gases, carbon dioxide.
Carbon dioxide, or CO2, is the primary greenhouse gas emitted by
human activities, and policies to control modern climate change frequently
focus on reducing our emissions of this gas. But constructing rational
climate change policies requires more than just knowing how much of it
humans are dumping into the atmosphere. It requires an understanding of
the carbon cycle – how carbon moves between the atmosphere, ocean,
land biosphere, and rocks on the Earth. This will help us understand what
happens to carbon dioxide after it is emitted into the atmosphere, which in
turn will help us understand the future trajectory of our climate.
148

5.1 Greenhouse gases and our
atmosphere’s composition
As we learned in Chapter 4, the greenhouse effect occurs because our
atmosphere is mostly transparent to visible photons but absorbs infrared
photons. It turns out that only a few of the components of our atmosphere
actually absorb infrared photons, and it is these greenhouse gases that are
responsible for the Earth’s greenhouse effect. In this section, I describe the
composition of our atmosphere, with a particular focus on greenhouse
gases.
Approximately 78 percent of the dry atmosphere (“dry atmosphere”
excludes water vapor)1 is made up of diatomic nitrogen or N2 – two
nitrogen atoms bound together. About 21 percent is diatomic oxygen or
O2, which is two oxygen atoms bound together; this is the part of the
atmosphere that we need to breathe to survive. Argon atoms make up
approximately 1 percent of our atmosphere. None of these three
constituents, which together make up more than 99 percent of the dry
atmosphere, absorbs infrared photons, so they are not greenhouse gases;
therefore, they do not warm the surface of the planet.
The next biggest component of the atmosphere is water vapor or H2O,
a constituent whose abundance varies widely from place to place. In the
warm tropics, water vapor can make up as much as 4 percent of the
atmosphere. In cold polar regions, in contrast, water vapor may be only 0.2
percent. Its abundance decreases rapidly with altitude, and in the
stratosphere it typically makes up 0.0005 percent of the atmosphere.
Water vapor is the most abundant and important greenhouse gas in
our atmosphere. Its main source is evaporation from the oceans, and it is
149

primarily removed from the atmosphere when water forms raindrops and
these fall to the surface. Emissions of water vapor from human activities
contribute essentially nothing to its atmospheric abundance. In Chapter 6, I
will talk in more detail about the role water vapor plays in climate change
and how humans are indirectly increasing its abundance.
Taken together, diatomic nitrogen and oxygen, water vapor, and
argon make up more than 99.95 percent of the atmosphere. You might
expect the remaining 0.05 percent to have no important role because it
seems like such a small amount, but you would be wrong. This last
smidgen of atmosphere is crucial to life on the planet.
The largest part of this remaining 0.05 percent is carbon dioxide or
CO2, which made up 0.04 percent of the atmosphere in 2014. Carbon
dioxide absorbs infrared photons and is therefore a greenhouse gas. In fact,
it is the second most important greenhouse gas, behind water vapor.
Because 0.04 percent is an awkwardly small number, scientists typically
express the concentration of these trace gases in a more convenient unit:
parts per million. A concentration expressed in parts per million indicates
how many molecules out of every million are the gas in question.2 In this
case, 0.04 percent corresponds to 400 parts per million or ppm, meaning
that there are 400 molecules of carbon dioxide in every million molecules
of air.
Parts per million can be usefully contrasted with percent, which
indicates how many molecules of every 100 are the species in question. In
fact, the word percent comes from the marriage of the words per cent,
literally meaning “out of 100.” Thus, air is approximately 78 percent
nitrogen, which means that 78 out of every 100 molecules of air are
molecules of nitrogen.
The next most important greenhouse gas in our atmosphere is
methane or CH4. In 2014, it had an atmospheric abundance of 1.83 ppm.
150

Despite its small abundance, methane is also a key player in our climate; I
will discuss it in detail in Section 5.7.
Another important greenhouse gas in our atmosphere is nitrous oxide
or N2O, which is present in today’s atmosphere at concentrations of about
0.32 ppm. This molecule is also known as “laughing gas,” which your
dentist might give you before she works on your teeth. It is emitted into
the atmosphere from nitrogen-based fertilizer and industrial processes as
well as several natural sources.
Ozone is another greenhouse gas. Its chemical formula is O3, so it is a
molecule made up of three oxygen atoms. The abundance of ozone varies
widely across the atmosphere – in unpolluted air near the surface, its
abundance is about 10-40 parts per billion,3 whereas its abundance can
reach 10 ppm in the stratosphere – 1,000 times higher.
Ozone is absolutely necessary for life on our planet because it absorbs
high-energy ultraviolet photons emitted by the Sun before they reach the
Earth’s surface. These photons carry enough energy that they can seriously
damage living tissue – leading to diseases such as skin cancer in humans.
But ozone is also one of the primary components of photochemical smog,
and breathing it can lead to health problems in humans and animals;
ground-level ozone can also damage plants. Thus, you want ozone
between yourself and the Sun, but you do not want to breathe it. Because
of this, ozone high up in the stratosphere is considered “good” ozone,
whereas ozone near the ground is “bad” ozone.
A final group of greenhouse gases are the halocarbons, including
chlorofluorocarbons and hydrochlorofluorocarbons, which are synthetic
industrial chemicals used as refrigerants (e.g., in air conditioners and
refrigerators) and in various industrial applications. This category also
includes natural molecules such as methyl chloride. Together, they are
present in today’s atmosphere at a concentration of a few parts per billion,
151

and all of them are powerful greenhouse gases. These halocarbons are also
the main culprits behind ozone depletion.
These greenhouse gases are not equal in their ability to warm the
planet. Methane is roughly twenty times more powerful than carbon
dioxide on a per molecule basis – meaning that it takes twenty molecules
or so of carbon dioxide to equal the warming from one molecule of
methane. The most powerful greenhouse gases on a per molecule basis are
the halocarbons. It takes several thousand carbon dioxide molecules to
equal the warming from one halocarbon molecule – so despite being about
1/10,000th as abundant as carbon dioxide, these halocarbons nevertheless
make an important contribution to the greenhouse effect. Do not forget,
too, that the three most abundant gases in our atmosphere, namely
diatomic nitrogen, diatomic oxygen, and argon, which together make up
about 99.9 percent of the dry atmosphere, are not greenhouse gases at all.
These differences in warming potential among various gases have
important implications when designing policies to address climate change.
As you will see, carbon dioxide is the most important greenhouse gas
for the problem of modern climate change. Because of this, this chapter
will mainly focus on it and the processes that regulate its atmospheric
abundance, which are collectively known as the carbon cycle.
152

5.2 Atmosphere-land biosphere-ocean
carbon exchange
153

(5.1)
5.2.1 Atmosphere-land biosphere exchange
We have been directly monitoring the abundance of carbon dioxide in the
atmosphere since the middle of the twentieth century. Figure 5.1 plots two
years (twenty-four months) of measurements, showing that the amount of
carbon dioxide in the atmosphere varies throughout the year: during its
maximum in May, carbon dioxide is about 6 ppm higher than the
September minimum.
Figure 5.1 The atmospheric abundance of carbon dioxide from fall
1988 through fall 1990
(data measured at Mauna Loa, Hawaii, and obtained from the NOAA
Earth System Research Laboratory/Global Monitoring Division; see
ftp://aftp.cmdl.noaa.gov/products/trends/co2/co2_mm_mlo.txt).
This annual cycle in carbon dioxide reflects the annual cycle of plant
growth and decay. Plants absorb carbon dioxide from the atmosphere and
use it to produce more plant material in a process known as
photosynthesis:
154

http://ftp://aftp.cmdl.noaa.gov/products/trends/co2/co2_mm_mlo.txt

(5.2)
In this reaction, carbon dioxide, water, and sunlight combine to produce
CH2O and O2. Because CH2O is a combination of carbon and water,
molecules made up of this unit are generally referred to as carbohydrates –
in this context, you can think of CH2O as the chemical formula for a plant.
Diatomic oxygen produced in this reaction is released into the atmosphere.
This is the main source for the oxygen in our atmosphere, which we
breathe to survive.
At the same time, humans, animals, and bacteria consume plant
material in order to produce energy through a reaction known as
respiration:
The net result of Equation 5.2 is carbon dioxide, which is released back
into the atmosphere, and energy, which is used to power the organism. It
should be noted that Equations 5.1 and 5.2 do not represent actual
chemical reactions; rather, they represent the net of a large number of
complex biochemical reactions that occur within the cells of organisms.
Equation 5.2 is the reverse of Equation 5.1: the carbon dioxide
consumed in the production of the plant material in Equation 5.1 is
released back into the atmosphere when the plant is consumed in Equation
5.2. Similarly, the oxygen molecule produced in Equation 5.1 is consumed
in Equation 5.2. The production of a carbohydrate through photosynthesis
followed by its consumption during respiration therefore produces no net
change in either carbon dioxide or oxygen. Instead, the net effect is the
conversion of sunlight into energy to power living creatures.
The atmosphere contained approximately 850 gigatonnes of carbon
(GtC) in 2014. A gigatonne is 1 billion metric tons, where 1 metric ton is
1,000 kg or 2,200 lbs. Note that this is just the mass of the carbon atoms in
the atmosphere – although the carbon dioxide molecule also contains two
155

oxygen atoms, their mass is not included. Unfortunately, you will also see
the mass expressed as the mass of carbon dioxide, which does include the
mass of the two oxygen atoms. In that case, the atmosphere contains more
than 3,100 GtCO2 (billion metric tons of carbon dioxide). You can convert
between these units by using the fact that 1 GtC = 3.67 GtCO2. In this
book, I will use GtC exclusively, but you must be careful to identify
whether the mass is given in GtC or GtCO2 when you read anything about
climate change.
The land biosphere contains 2,500 GtC, stored in living plants and
animals and in organic carbon in soils (e.g., decaying leaves). During a
given year, photosynthesis removes approximately 120 GtC from the
atmosphere. Respiration roughly balances this, transferring about the same
amount back to the atmosphere. Thus, over a year, there are only small
changes in carbon dioxide in the atmosphere or land biosphere as a result
of photosynthesis or respiration.
The fact that photosynthesis and respiration are balanced over the
year does not mean that they are in balance at every point in time. Most of
the Earth’s land area – and, therefore, most of the Earth’s plants – are found
in the northern hemisphere. During the northern hemisphere’s spring and
summer (May-September), when plants are growing and trees are leafing,
global photosynthesis exceeds respiration and there is a net drawdown of
carbon dioxide out of the atmosphere and into the land biosphere; we can
see this in Figure 5.1.
During the northern hemisphere’s fall and winter (October-April),
plant material that was produced during the spring and summer decays,
releasing carbon dioxide back into the atmosphere. During this period,
global respiration exceeds photosynthesis and there is a net transfer of
carbon from the biosphere into the atmosphere, which we can also see in
Figure 5.1.
156

There is also a large amount of carbon stored in permafrost, which is
ground that is frozen year-round. Much like that frozen dinner that has
been in your freezer since Bill Clinton was President, dead organic plant
matter frozen into the permafrost does not decay; it is kept intact as long as
the ground remains frozen. If permafrost thaws out, however, the organic
matter stored there will begin to decay, releasing carbon into the
atmosphere.
While permafrost is indeed melting, it appears at present that this
contributes little to atmospheric carbon dioxide. However, given that much
of this permafrost is in the Arctic, which is expected to continue warming
rapidly, the melting will continue. At some point, perhaps soon, the
resulting release of carbon dioxide may contribute significantly to
atmospheric carbon dioxide.
An aside: Where does the oxygen in our atmosphere come from?
As I mentioned earlier, photosynthesis followed by respiration is
not a net producer or consumer of carbon dioxide or oxygen.
Where, then, does the large amount of molecular oxygen in our
atmosphere come from? It turns out that it is the result of
photosynthesis that is not balanced by respiration. That occurs
when a plant grows through photosynthesis, but the plant material
is buried before it can be consumed via respiration. When that
happens, the oxygen produced during photosynthesis is not
consumed. Over the billions of years that life has existed on the
planet, this process has built up and now maintains the oxygen
levels in our atmosphere.
157

(5.3)
(5.4)
5.2.2 Atmosphere-ocean carbon exchange
One of carbon dioxide’s most important properties is that it readily
dissolves in water. Once it has dissolved in water, carbon dioxide is
converted to carbonic acid (H2CO3) by means of this reaction:
This process is sometimes referred to as ocean acidification. As the oceans
become more acidic, the biology of the oceans can change – and given
human reliance on the oceans for food, this could lead to important
impacts on humans. This will be discussed in more detail in Chapter 9.
The carbonic acid formed in Equation 5.3 can react further with water
to convert into other forms of carbon. Because of the conversion of carbon
dioxide to many other forms of carbon, the ocean can absorb huge
amounts of carbon dioxide. Carbon is returned to the atmosphere in a
reaction that is the reverse of Equation 5.3:
This is followed by the escape of carbon dioxide back into the atmosphere.
Processes embodied by Equations 5.2 and 5.3 transfer about 80 GtC per
year between the atmosphere and ocean, roughly similar to the exchange
between atmosphere and land biosphere.
Thus, carbon cycles easily between the atmosphere and ocean. To
fully understand this exchange, however, we must think of the ocean as
being split into two parts. The first part is the top 100 m or so of the ocean,
which exchanges carbon very rapidly with the atmosphere. This part of the
ocean makes up only a few percent of the mass of the ocean and is often
158

referred to as the mixed layer because it is well mixed by winds and other
weather events; it contains 900 GtC.
Below this lies the other 97 percent of the ocean, known as the deep
ocean. The deep ocean also contains most of the ocean’s carbon,
approximately 40,000 GtC, or forty-seven times more carbon than is in the
atmosphere. The mixed layer and the deep ocean exchange carbon at a rate
of about 100 GtC per year. This occurs as ocean currents mix high-carbon
water from the mixed layer with low-carbon water from the deep ocean. It
also occurs when sinking organic matter, such as dead organisms or fecal
material, falls from the mixed layer into the deep ocean – a process known
as the biological carbon pump.
159

5.2.3 The combined atmosphere-land biosphere-ocean system
Figure 5.2 shows a schematic of the combined atmosphere-land biosphere-
ocean system. Approximately 120 GtC per year are continuously cycling
between the atmosphere and land biosphere as plants absorb carbon
dioxide as they grow and then release carbon dioxide when they die.
About 80 GtC per year of carbon from the atmosphere is continuously
dissolving into the ocean’s mixed layer, while the same mass of carbon
atoms is coming out of the ocean and back into the atmosphere, thereby
cycling between the atmosphere and ocean. At the same time, the mixed-
layer and deep ocean are exchanging 100 GtC per year.
Figure 5.2 A schematic of exchange between the atmosphere, land
biosphere, and ocean. Reservoirs are in GtC; fluxes are in GtC per year.
Based on Figure 6.1 of Ciais et al. [2013].
160

To get an idea of what these numbers actually mean, we calculate
turnover times for the atmosphere and land biosphere reservoirs. The
turnover time for the atmosphere is the length of time that a carbon atom in
the atmosphere will remain there before being transferred into one of the
other two reservoirs. This can be roughly estimated as the size of the
reservoir, 850 GtC, divided by the total flux out of the reservoir, 200 GtC
per year (120 GtC per year goes into the land biosphere and 80 GtC per
year goes into the mixed layer). This yields an atmospheric turnover time
of about four years. This turnover time is also referred to as a “lifetime” or
“residence time.”
This means that a carbon atom stays in the atmosphere for only four
years or so before it is transferred into the land biosphere or ocean.
Remember that this is an average value – an individual molecule of carbon
dioxide may remain in the atmosphere for a shorter or longer time.
Another way to think about a turnover time is that, over a period of four
years, enough exchange will take place to replace all of the carbon that is
in the atmosphere with carbon from the land biosphere or ocean.
The turnover time of carbon in the land biosphere is 2,500 GtC
divided by 120 GtC per year = 21 years. This means that a carbon atom in
the land biosphere will stay there for 21 years before being transferred into
the atmosphere. Thus, it takes a few decades for a carbon atom to make a
round trip from the land biosphere into the atmosphere and back into the
land biosphere.
We can also calculate the turnover times for the ocean reservoirs. The
total flux out of the mixed layer is 180 GtC per year (80 GtC per year is
exchanged with the atmosphere and 100 GtC per year is exchanged with
the deep ocean), so the turnover time for the mixed layer is 900 GtC ÷ 180
GtC per year ≈ five years. The turnover time for the deep ocean is several
centuries: 40,000 GtC ÷ 100 GtCr ≈ 400 years. Thus, it takes a few
161

centuries for a carbon atom to make a round trip from the atmosphere
through the mixed layer, the deep ocean, and back.
Another way to think about this is that the atmosphere exchanges
carbon rapidly (time scale of years to decades) with the land biosphere and
mixed layer, and much more slowly (time scale of centuries) with the deep
ocean. Later in the chapter, I will explain why this is so important for the
climate change problem.
162

5.3 Atmosphere-rock exchange
Most of the carbon in the world – many millions of gigatons of carbon – is
stored in rocks, such as limestone (CaCO3), and this carbon is slowly
exchanging with the atmosphere-land biosphere-ocean system (Figure
5.3). Carbon dioxide is transferred from rocks directly into the atmosphere
by volcanic eruptions. This process releases an average of 0.1 GtC per
year. Although this flux is small compared with other fluxes, over millions
of years it can lead to significant transfers of carbon into the atmosphere-
land biosphere-ocean system.
Figure 5.3 A schematic of exchange between the atmosphere-land
biosphere-ocean reservoir and the rock reservoir (reservoirs are in GtC;
fluxes are in GtC per year).
These natural emissions of carbon dioxide from the rock reservoir are
roughly balanced by a process known as chemical weathering, which
removes about an equal amount of carbon from the atmosphere and
transfers it back into rocks. Chemical weathering starts when carbon
dioxide in the atmosphere dissolves into raindrops falling toward the
surface (remember that carbon dioxide dissolves readily in water).
Carbonic acid (H2CO3) is produced via Equation 5.3, which makes the rain
slightly acidic (pH = 5.6).
163

(5.5)
When this acidic rain falls on rocks, both the physical impact of the
rain and chemical reactions break the rock down. The chemical reaction is
shown here:
Note that this equation is a general description of the process of
weathering, not the exact chemical reaction. Nonetheless, the essential
message of Equation 5.5 is correct: The carbon dioxide molecule
consumed in this reaction came from the atmosphere, via rainwater, and it
is transferred into a molecule of calcium carbonate or CaCO3, which is
limestone, and subsequently runs off with the rainwater and eventually
reaches the ocean. The reaction also forms silicon dioxide (SiO2), the
primary component of sand, quartz, and glass.
Once in the ocean, the molecules of calcium carbonate are deposited
through various mechanisms on the sea floor. Over many millions of
years, plate tectonics carries this calcium carbonate deep within the Earth,
where high temperatures and pressures turn the rock into magma.
Eventually, this carbon is transferred back to the surface by volcanism,
thereby releasing the carbon dioxide back to the atmosphere and
completing the cycle.
Another pathway for carbon to move into the rock reservoir occurs
when plants are rapidly buried in sediment before they can decay. This is,
in fact, the same process that leads to a net production of oxygen
(discussed earlier). Once buried and subjected to the great heat and
pressure found deep within the Earth, this dead plant material can be
converted to fossil fuels, which humans extract and burn for energy –
thereby returning the carbon to the atmosphere. We will explore fossil
fuels in the next section.
164

A carbon atom will remain in the atmosphere-land biosphere-ocean
system for approximately 46,000 GtC ÷ (0.1 GtC per year) = 460,000
years before it is transferred into the rock reservoir. Given the large size of
the rock reservoir and the relatively small rate of exchange between the
rocks and the atmosphere, it takes many, many millions of years for a
carbon atom to travel through the rock reservoir and reemerge into the
atmosphere.
Figure 5.4 shows an estimate of atmospheric carbon dioxide over the
past half-billion years. Four hundred million years ago, atmospheric
carbon dioxide was more than ten times higher than it is today, and since
that time it has generally decreased, although there have been wide
variations. The abundance in 2014, 400 ppm, is relatively low when
compared to the geologic record (although it is higher than it has been for
several million years).
Figure 5.4 Atmospheric carbon dioxide over the past half-billion years
(based on Royer, 2006, Figure 1).
The variations in atmospheric carbon dioxide in Figure 5.4 are due to
variations in the rate of exchange of atmospheric carbon with the rock
reservoir. This includes variations in the rate at which carbon dioxide is
emitted from volcanoes – during periods of extreme volcanism, for
example, atmospheric carbon dioxide will increase.
165

The movement of the continents is another factor. As continents
move, patterns of rainfall can change, and new rock can be exposed to the
atmosphere, both of which can change the rate of chemical weathering –
and therefore the rate at which carbon dioxide is removed from the
atmosphere. For example, approximately 40 million years ago, the Indian
subcontinent collided with the Asian continent, forming the Himalayas and
the adjacent Tibetan Plateau. Changing wind patterns brought heavy
rainfall onto the expanse of newly exposed rock, and the resultant
chemical weathering has been drawing down atmospheric carbon dioxide
since then – until humans came along, that is.
166

5.4 How are humans perturbing the
carbon cycle?
As Figure 5.4 shows, carbon dioxide varies without any human activities.
However, humans can also affect the carbon cycle. Figure 5.5 shows the
perturbed carbon cycle, with the flows of carbon caused by human
activities indicated as the gray lines. The main perturbation comes from
the combustion of fossil fuels for energy. Fossil fuels were formed when
plants that grew hundreds of millions of years ago were buried before the
carbon in them could be released back into the atmosphere by respiration.
Under high pressure and heat, applied over millions of years, the carbon in
the plants was converted into the substances we know today as oil, coal,
and natural gas.
Figure 5.5 Diagram of the carbon cycle as perturbed by humans. Gray
arrows show net flows of carbon caused by human activities. Arrows B,
C, and D represent deforestation, enhanced absorption of carbon by the
land biosphere, and enhanced absorption of carbon by the ocean,
respectively.
When fossil fuels are burned, the net reaction is similar to the
respiration reaction (Equation 5.2):
167

(5.6)Fossil fuels are represented in Equation 5.6 by CHx
because they are primarily carbon, with varying amounts of hydrogen.
During combustion, the fossil fuel combines with oxygen to produce
energy, carbon dioxide, and an amount of water vapor that depends on
how much hydrogen was in the fuel. The resulting energy is used to power
our world, and the carbon dioxide is vented directly into the atmosphere.
Fossil fuels can also contain other trace species, such as sulfur. When
burned, these trace species can also be released into the environment and
lead to environmental problems of their own, such as acid rain. Also note
that Equation 5.6 is a schematic reaction, not an actual chemical reaction,
so do not be concerned that it does not balance.
Before humans discovered them, fossil fuels were safely sequestered
in the rock reservoir. The natural carbon cycle would have slowly released
this carbon back to the atmosphere through geologic processes over many
millions of years. Humans, however, are extracting and combusting fossil
fuels at a breath-taking pace – fast enough that we will extract and burn
most of the world’s fossil fuels in just a few hundred years. The net result
is the creation of an additional pathway for carbon from rocks to the
atmosphere (the line marked “Fossil fuels” in Figure 5.5). And this
pathway is large: over the period 2002–2011, the combustion of fossil
fuels led to average emissions of approximately 8.3 GtC per year to the
atmosphere.4 This is more than eighty times the natural flow rate of carbon
from the rock reservoir to the atmosphere.
Humans have also been chopping down large tracts of forest – a
process known as deforestation – in order to use the land for other
activities, such as agriculture or grazing livestock. Frequently, the forest is
removed by burning it, which releases the carbon stored in trees and other
plants to the atmosphere. Even just bulldozing the forest releases the
carbon to the atmosphere, albeit more slowly. In addition, the soil often
168

contains large amounts of carbon stored in the form of dead organic plant
material. When the forest is removed, much of this plant material
decomposes following Equation 5.2, releasing the carbon back into the
atmosphere.
Deforestation is just one of many ways that man’s impact on the land
can influence atmospheric carbon dioxide. Emissions associated with these
changes are known collectively as land-use changes. Land-use changes are
an important source of carbon dioxide for the atmosphere, and estimates
are that it contributed approximately 0.9 GtC per year to the atmosphere
from 2002 to 2011 – about one-tenth of the emissions from fossil fuel
combustion. This flux is shown in Figure 5.5 as Arrow B.
Figure 5.6a shows that the abundance of carbon dioxide in our
atmosphere has remained in a narrow range, 260–280 ppm, over the last
10,000 years – until a sudden spike occurred in the past few hundred years.
Figure 5.6b focuses on this spike, showing that atmospheric carbon
dioxide began rapidly increasing around 1800. This coincides with the
industrial revolution, when widespread burning of fossil fuels began.
169

Figure 5.6 Abundance of carbon dioxide in our atmosphere over the
past (a) 10,000 years, (b) 250 years, and (c) 50 years (sawtooth line).
Panel c also shows (gray curve) the annual average carbon dioxide
abundance if all of the carbon dioxide emitted by human activities since
1959 had remained in the atmosphere.
Panels a and b are adapted from Figure SPM.1 of IPCC, 2007a;
measurements in panel c are obtained from the NOAA Earth System
Research Laboratory/Global Monitoring Division, at
http://www.esrl.noaa.gov/gmd/ccgg/trends/; the 100 percent-retained
line is calculated by assuming that 55 percent of the emitted carbon is
rapidly removed.
Figure 5.6b also shows that the rise in carbon dioxide is accelerating.
Over the past 250 years, atmospheric carbon dioxide has increased by
about 120 ppm. The first 60 ppm of the increase took more than 200 years,
until 1980, whereas the next 60 ppm took only thirty-five years. Figure
170

http://www.esrl.noaa.gov/gmd/ccgg/trends/

5.6c shows high-resolution measurements of the abundance of carbon
dioxide in our atmosphere over the past fifty-five years. The yearly
sawtooth pattern reflects the seasonal cycle in plant growth, which was
discussed in Section 5.2.1. There is also a long-term increase in carbon
dioxide, from 315 ppm in the late 1950s to 400 ppm in 2014, caused by
fossil fuel combustion and land-use changes.
Figure 5.7 shows the year-to-year increase in atmospheric carbon
dioxide. In the late 1950s, the increase in atmospheric carbon dioxide was
less than 1 ppm per year. By the first decade of the twenty-first century,
atmospheric carbon dioxide was increasing by nearly 2 ppm per year. This
reflects the increasing emissions due to increasing fossil fuel combustion
over the past half-century.
Figure 5.7 Bars show the observed year-to-year increase in atmospheric
carbon dioxide. The solid line shows what the annual increase would
have been had 100 percent of the carbon dioxide emissions remained in
the atmosphere
(bars are provided by the NOAA Earth System Research
Laboratory/Global Monitoring Division, at
http://www.esrl.noaa.gov/gmd/ccgg/trends/; the solid line is from
Denman et al., 2007, Figure 7.4).
171

http://www.esrl.noaa.gov/gmd/ccgg/trends/

We are emitting carbon dioxide to the atmosphere, and we can see
that the amount of carbon dioxide in the atmosphere is increasing. So far,
so good. But there is a problem. We have good records of exactly how
much fossil fuel is extracted and burned each year, so we can calculate
how much atmospheric carbon dioxide should have increased. During
2002–2011, human emissions of carbon to the atmosphere averaged 9.2
GtC per year. But during that time, the increase in atmospheric carbon
dioxide averaged 4.3 GtC per year – thus, the increase in carbon dioxide
was about half of what we were emitting.
Figure 5.7 shows this more clearly: the solid curve is an estimate of
how much atmospheric carbon dioxide should have increased if it had all
stayed in the atmosphere. Over the past half century, the increase in
atmospheric carbon dioxide each year is approximately half of the amount
emitted.
So where are the rest of our emissions going? It turns out that about
half of this “missing carbon” is dissolving into the ocean. This is indicated
as Arrow D in Figure 5.5. The other half is believed to have gone into the
land biosphere, although there is considerable scientific debate about
exactly what part of the land biosphere is absorbing the carbon. This
enhanced land sink is represented by Arrow C in Figure 5.5.
An aside: How do we measure how much carbon is going into the
ocean versus the land-biosphere?
The curve of observed atmospheric carbon dioxide abundances
(Figure 5.6c) is often referred to as the Keeling Curve, named after
Charles D. Keeling, the scientist who initiated the measurements in
1957. The acorn, as they say, never falls far from the tree, and
Keeling’s son, Ralph, is following in his father’s footsteps by
172

making long-term measurements of another key atmospheric
measurements: diatomic oxygen, O2.
Figure 5.8 shows Ralph Keeling’s measurements of oxygen
over the last twenty-five years, along with measurements of carbon
dioxide. The annual sawtooth, due to the seasonal cycle of plant
growth, is apparent in both time series. However, the plots are
exactly out of phase – for example, when the carbon dioxide is
going up, oxygen is going down. Equations 5.1 and 5.2 explain
why this is so: when carbon dioxide is absorbed by a plant during
photosynthesis, oxygen is released to the atmosphere; when carbon
dioxide is released during respiration, oxygen is absorbed from the
atmosphere.
Figure 5.8 Time series of the change (in percent) of the O2/N2
ratio (relative to a standard) and CO2 mixing ratio (in ppm) at the
Canadian Alert Station (82°N). Diatomic nitrogen (N2) is not
changing much, so most of the change in the ratio is due to
changes in oxygen. The main cause of the overall downward trend
in O2/N2 is the loss of O2 due to fossil fuel burning, which is also
the main cause of the overall rise in CO2.
173

Keeling’s data also show a long-term decline of oxygen in the
atmosphere. This decline is the result of burning fossil fuels (see
Equation 5.6); given the mix of fuels we use, about 1.4 molecules
of oxygen are consumed for every carbon dioxide molecule
released. So the good news is that we finally found a gas in the
atmosphere whose abundance is declining – the bad news is that it
is oxygen!5
The exact amount of oxygen lost gives us important
information about how much carbon the land and ocean are
absorbing because the different pathways lead to different changes
in oxygen. To see how this works, imagine that some fossil fuel is
burned, and this adds 100 molecules of carbon dioxide to our
atmosphere; this would also remove 140 molecules of oxygen.
As discussed earlier, about half of the carbon dioxide (in this
case, fifty molecules) is absorbed by some combination of the land
biosphere and ocean. Absorption by the ocean does not affect
oxygen (Equation 5.3), so combustion followed by ocean
absorption would lead to a net decrease in oxygen of 140
molecules.
Absorption by the land biosphere, however, produces oxygen
(Equation 5.1) – about 1.1 molecules of oxygen are produced for
every carbon dioxide molecule that is absorbed during
photosynthesis. If all fifty carbon dioxide molecules are absorbed
by the land biosphere, then 50 × 1.1 = 55 molecules of oxygen
would be produced. So the net change in oxygen in this case would
be a net loss of 140 − 55 = 95 molecules of oxygen.
From Keeling’s oxygen measurements, we find that the
decline in oxygen falls about midway between these limits, leading
174

to the conclusion that about half of the carbon is going into the
land biosphere and half into the ocean.
Figure 5.7 also shows large variability in the year-to-year increase of
atmospheric carbon dioxide. For example, atmospheric carbon dioxide
increased by nearly 3 ppm in 1998, but less than 1 ppm the very next year
– even though emissions were roughly the same. This is mainly due to
variations in the climate due to El Niño events. Regional climate variations
during these events modify the uptake and emissions of carbon from the
land biosphere by varying areas of rainfall and drought.
If the land biosphere and ocean were not taking up about half of the
carbon we emit, then atmospheric carbon dioxide would be much higher
today than it actually is. Figure 5.6c also shows a rough estimate of what
the long-term time series of atmospheric carbon dioxide would be if all
emissions had remained in the atmosphere – it shows that atmospheric
carbon dioxide might be near 500 ppm in 2014, approximately 100 ppm
higher than the actual abundance. The climate would consequently be
warmer and changing even more rapidly. Thus, the land biosphere and
ocean are doing us a huge favor by absorbing significant amounts of
carbon emitted by humans.
An emerging concern is whether the oceans and land biosphere can
continue taking up as much carbon in the future as they presently are. It is
unknown when or if we will reach a saturation point at which point the
reservoirs slow down or even cease their uptake. If that happens, then a
higher fraction of emissions will remain in the atmosphere, and the
abundance of carbon dioxide in our atmosphere will grow more rapidly.
This leads to yet another worry for climate scientists: that climate change
175

itself may alter the carbon cycle. In the next chapter, we will explore how
such carbon cycle feedbacks may amplify climate change.
176

5.5 Some commonly asked questions
about the carbon cycle
Because of its central role in the climate change problem, climate skeptics
occasionally challenge the claim that the observed increase in carbon
dioxide since 1800 is due to human activities. In this section I address this
argument.
How do we know that combustion of fossil fuels is responsible for the
increase in carbon dioxide, rather than nonhuman sources such as
volcanoes or plants?
This is a reasonable question. After all, the amount of carbon dioxide
absorbed by plants during the year and balanced by plant decay
(approximately 120 GtC per year) is much larger than human emissions
(which averaged 9 GtC per year over 2002–2011) – ditto for the ocean
fluxes. So it may seem reasonable that the increase in atmospheric carbon
dioxide might be driven by a slight excess of plant respiration over
photosynthesis, or a slight excess flux of carbon dioxide out of the ocean.
Similarly, we know that volcanoes emit carbon dioxide, and that over
millions of years, volcanoes are a primary source of carbon dioxide in the
atmosphere. So maybe the increase in atmospheric carbon dioxide is due to
enhanced volcanic activity.
There are, however, several independent lines of evidence that
unanimously agree that fossil fuel combustion is the dominant reason for
the increase in atmospheric carbon dioxide over the past few centuries.
First, Figure 5.7 shows that, for the past half-century, each year’s increase
in carbon dioxide in the atmosphere has been on average about half of
177

what humans released into the atmosphere in that same year. Thus, when
humans were emitting smaller amounts of carbon dioxide in the 1960s,
atmospheric carbon dioxide was increasing at a slower rate than when
humans were dumping large amounts of carbon dioxide in the atmosphere,
as we are today. If the source of carbon dioxide emissions were nonhuman,
it seems unlikely that it would track human emissions of carbon dioxide so
closely.
Second, the carbon dioxide can be chemically “fingerprinted” to show
that it comes from fossil fuels. The method is based on isotopes of carbon.
All carbon atoms have six protons, but carbon’s isotopes have different
numbers of neutrons. The most abundant isotope is carbon-12, containing
six neutrons to go with the six protons, and which makes up roughly 99
percent of the carbon on Earth. Carbon-13, with seven neutrons, makes up
1 percent of the carbon, and approximately one carbon atom out of a
trillion is carbon-14, which has eight neutrons.
The chemical properties of an atom are for the most part set by the
number of protons, so isotopes tend to have very similar chemical
properties. The chemistry, though, is not identical. Plants, for example,
preferentially absorb carbon-12 when growing. And because fossil fuels
are derived from plants, they reflect this preference for carbon-12. When
the fossil fuels are burned, the carbon dioxide produced also reflects
plants’ preference for carbon-12 over carbon-13.
Scientists can measure the amount of carbon-12 and carbon-13 in
atmospheric carbon dioxide, and those measurements show that the
increase in atmospheric carbon dioxide in Figure 5.6b is caused by carbon
that is depleted in carbon-13 – such as that which comes from plants. This
allows us to rule out sources such as volcanoes or the ocean. The
measurements of atmospheric oxygen discussed earlier are also
inconsistent with a volcanic or oceanic source.
178

Thus, we know that the increase in atmospheric carbon dioxide is
coming from plants, but is it coming from plants that died hundreds of
millions of years ago (i.e., fossil fuels) or plants of today? In order to make
that determination, we turn to carbon-14. Carbon-14 is produced in the
atmosphere when a neutron created by a cosmic ray hits the nucleus of an
atom of nitrogen-14. The nucleus absorbs the neutron and ejects a proton,
thereby transforming itself into carbon-14. Carbon-14 atoms are
incorporated into molecules of carbon dioxide and are then absorbed by
plants and incorporated into plant material. If you walk outside and pull a
leaf off a tree, a small fraction of atoms in that leaf would be carbon-14.
Carbon-14 is known as radiocarbon because it is radioactive. That
means its nucleus is unstable and converts back to nitrogen-14 with a half-
life of approximately 6,000 years (so that, after 6,000 years, half of the
carbon-14 has converted back to nitrogen-14). To see the implications of
this, imagine a cotton plant that grew 6,000 years ago. As it grew, the plant
absorbed carbon dioxide containing carbon-14 from the atmosphere, and
the carbon-14 was incorporated into the plant. Immediately after the plant
was picked, it would have the same proportion of carbon-14 as any living
plant, and so would the cotton produced from it. But because it was no
longer alive, it stopped absorbing carbon dioxide from the atmosphere.
Over time, the amount of carbon-14 in the cotton slowly decreased as it
was converted back to nitrogen-14.
Now imagine that modern-day archaeologists find a blanket made of
this cotton and want to know how old it is. To do this, they measure the
proportion of carbon-14 in the blanket and find that it has half the carbon-
14 of a living plant. With a half-life of 6,000 years, the archaeologists
conclude that the blanket is 6,000 years old. If they found that it had one
fourth of the carbon-14 of a living plant, then it would be 12,000 years old.
This process is known as radiocarbon dating.
179

Now let us turn our attention to fossil fuels. As we learned earlier,
fossil fuels are produced when plant matter is buried for millions of years.
After millions of years of being underground, all of the carbon-14 has
converted back to nitrogen-14. Thus, fossil fuels contain essentially no
carbon-14, a condition known as radiocarbon dead. So when the fossil
fuels are burned, the carbon dioxide produced also has no carbon-14 in it.
Scientists measuring the isotopic composition of atmospheric carbon
dioxide have found that the carbon dioxide being added to the atmosphere
is indeed radiocarbon dead, showing that it is coming from long-dead
plants – fossil fuels – and not modern plants.
Putting all of the evidence together, along with an absence of any
counterevidence, we see that there is no question that human activities are
increasing the amount of carbon dioxide in the atmosphere. As my
colleague John Nielsen-Gammon puts it, not only can we see the smoking
gun, but the smoke is a chemical match to the gunpowder.
Why focus on carbon dioxide from fossil fuel combustion when
plants and animals emit far more carbon dioxide to the atmosphere?
Humans, animals, bacteria, and plants do indeed emit enormous amounts
of carbon dioxide to the atmosphere – the land biosphere emits 120 GtC
per year, compared with present-day emissions from human activities of
about 9 GtC per year. So why should we care about carbon dioxide from
fossil fuels? To understand the answer, you need to understand the
difference between carbon dioxide coming from fossil fuel combustion
and from respiration by living organisms.
Let us begin by imagining that you plant a carrot seed, and over the
next few months this seed grows into a carrot. As described by Equation
5.1, the plant grows by absorbing carbon dioxide directly from the
180

atmosphere, and this reduces the amount of carbon dioxide in the
atmosphere. Now let us imagine that the carrot is eaten by a goat. The goat
metabolizes the carrot (in a manner approximately following Equation
5.2), which produces energy to power the goat’s vital functions. The
carbon dioxide produced is exhaled back into the atmosphere.
Thus, when an animal exhales carbon dioxide, it is releasing back into
the atmosphere carbon dioxide that was in the atmosphere just a few
months before. Although this can lead to seasonal variations in carbon
dioxide, as shown in Figure 5.1, it does not cause long-term increases in
carbon dioxide. Figure 5.6a confirms this by showing basically no change
in carbon dioxide over the past 10,000 years – a period during which
humans, plants, and animals were certainly releasing carbon dioxide to the
atmosphere.
In contrast, when you burn fossil fuels, you are releasing to the
atmosphere the carbon dioxide that had been safely sequestered in rocks
(e.g., Figure 5.5) for hundreds of millions of years. This is a net addition to
the atmosphere, so it does cause a long-term increase in carbon dioxide.
Figure 5.6b confirms this: The increase in atmospheric carbon dioxide
started with the industrial revolution, when society-wide burning of fossil
fuels began.
181

5.6 The long-term fate of carbon dioxide
To get a feel for the long-term evolution of the climate over the next
millennium, we need to know how long the carbon dioxide we release
stays in the atmosphere. As a thought experiment, imagine that a pulse of
carbon dioxide is instantaneously released into the atmosphere. At first, the
land biosphere and mixed-layer of the ocean rapidly take carbon dioxide
out of the atmosphere – Figure 5.9 shows that about 40 percent of the
carbon dioxide pulse is removed in twenty years. Removing additional
carbon dioxide requires transport into the deep ocean, which is a slower
process. After 400 years, about 25 percent of the slug of carbon dioxide
pulse remains in the atmosphere.
Figure 5.9 Fraction of carbon dioxide remaining in the atmosphere after
an initial pulse in year zero. The plot shows that it takes a very long time
for carbon dioxide emitted to the atmosphere to be completely removed.
Based on Figure 1 of Box 6.1 of Ciais et al. (2013).
At this point, the deep ocean is in equilibrium with the atmosphere
and cannot absorb any more carbon. Further removal of carbon dioxide
requires reactions between carbon dissolved in the ocean and calcium
carbonate (CaCO3) sea floor sediments. These reactions transfer carbon to
the ocean sediments, allowing the ocean to absorb more carbon dioxide.
182

But this process is very slow – after 10,000 years, 15 percent of the initial
pulse of carbon dioxide is still in the atmosphere. The last 15 percent is
removed by chemical weathering (Equation 5.5) over the next few hundred
thousand years.
The very long time it takes for carbon to be removed from the
atmosphere is confirmed by estimates of carbon dioxide during the
Paleocene-Eocene Thermal Maximum (discussed in Section 2.2.2; see
Figure 2.11). This is an event about 55 million years ago when a huge
pulse of carbon (several thousand GtC) was released into the atmosphere,
leading to a sudden and significant warming of the planet. It took several
hundred thousand years for that slug of carbon to be removed and for the
warming it caused to dissipate.
The upshot is that, if we add carbon dioxide to the atmosphere, it will
remain in the atmosphere for a very long time, and the warming it causes
therefore also sticks around for a very long time. We will revisit the grim
implications of this in Chapter 8, but it is useful to realize that the actions
we take in the next few decades will determine the trajectory of the climate
for thousands, if not tens of thousands of years. If our actions this century
lead to massive, long-term climate change, I wonder what people living in
much warmer climates in the years 3000, 4000, or 10,000 will think of us.
183

5.7 Methane
Most discussions of the carbon cycle focus on the cycling of atmospheric
carbon dioxide. However, methane is another crucial carbon-containing
gas. Although the atmospheric abundance of methane was only 1.83 ppm
in 2014, much smaller than that of carbon dioxide, on a per molecule basis,
methane is roughly twenty times more powerful a greenhouse gas than
carbon dioxide.
Methane is emitted to the atmosphere from both human and natural
processes. About 60 percent of human methane emissions are from
agriculture and waste. Livestock is the largest source of methane in this
category. Cattle, as well as goats and sheep, are ruminants, and these
animals produce methane in their guts during the digestion of food. This
methane is eventually released to the atmosphere (out of both ends of the
animals). The next largest source in this category is bacterial processes in
landfills and other waste repositories. Emissions from rice paddies are the
third significant source. In the warm and wet environment of a flooded rice
field, bacteria in the soil efficiently produce methane, the vast majority of
which is then released to the atmosphere.
The release of methane from the petrochemical industry is
responsible for about 30 percent of human emissions of methane. This
comes from leakage of methane from natural gas wells as well as release
of geologic methane from coalmines. Finally, burning of forest and other
biomass primarily produces carbon dioxide, but it also produces methane
if the combustion temperature is sufficiently low (e.g., a smoldering fire).
This is responsible for the remaining 10 percent of human methane
emissions.
184

During the 2000s, methane emissions from natural sources were
about equal to human emissions. Approximately two-thirds of these
natural emissions were from natural wetlands, which produce methane the
same way that flooded rice paddies do. Minor contributions come from the
ocean, from freshwater lakes and rivers, and from wild animals,
particularly termites.
Figure 5.10 shows that the atmospheric abundances of methane began
rising about 1800, the same point at which carbon dioxide began rising.
Scientists have also attributed this increase to human activities. As we will
see in Chapter 6, these emissions are enough to make methane a
significant contributor to the warming we are experiencing – methane’s
contribution to global warming is approximately one-fourth of the
contribution of carbon dioxide. As a result, reductions of methane
emissions are frequently included in plans to address climate change.
Figure 5.10 Methane abundance over the past 10,000 years. The inset
plot shows a close-up of the past 250 years
(adapted from the IPCC, 2007a, Figure SPM.1).
Methane is removed from the atmosphere by oxidation, which
follows the following schematic reaction:
185

On average, a molecule of methane is destroyed by this reaction ten years
after it was emitted. If we stopped emitting methane today, within a few
decades, all of the human-emitted methane would be gone, and the
atmospheric abundance would be back down to preindustrial amounts.
This is quite different from carbon dioxide, which can stay in the
atmosphere for centuries or millennia.
186

5.8 Chapter summary
Only a few components of our atmosphere are greenhouse gases,
which absorb infrared photons. The three most important are (in
order) water vapor, carbon dioxide, and methane. Nitrogen,
oxygen, and argon, which make up approximately 99.9 percent of
the dry atmosphere, are not greenhouse gases.
The carbon cycle describes how carbon cycles through its primary
reservoirs: the atmosphere (containing 850 GtC), land biosphere
(2,500 GtC), ocean (900 GtC in the mixed layer and 40,000 GtC in
the deep ocean), and rocks (millions and millions of GtC).
The atmosphere exchanges carbon with the land biosphere through
photosynthesis and respiration. The atmosphere exchanges carbon
with the ocean when carbon dioxide dissolves into or is emitted
from the ocean. Once in the ocean, the carbon dioxide is converted
to carbonic acid and other chemicals. Over the course of several
centuries, a carbon atom added to the atmosphere will cycle
through all of the other reservoirs and return to the atmosphere.
The atmosphere-land biosphere-ocean system also exchanges
carbon with rock reservoirs through volcanism and chemical
weathering. This exchange is extremely slow.
Humans are perturbing the carbon cycle by extracting and burning
fossil fuels. The result is the creation of a new, rapid pathway
moving carbon from rocks to the atmosphere. Between 2002 and
2011, fossil fuel combustion released an average of 8.3 GtC to the
atmosphere from the rock reservoir, which is more than eighty
187

times the amount released from the rocks naturally. Land-use
change is another important human source, releasing 0.9 GtC per
year from the land-biosphere into the atmosphere during this
period.
This has increased atmospheric carbon dioxide abundance from
approximately 280 ppm in 1750, before the industrial revolution, to
400 ppm in 2014.
It takes a long time for the carbon cycle to remove carbon that
humans add to the atmosphere. About 40 percent of it is removed
in a few decades, 75 percent in a few centuries, and the last 25
percent is removed over tens and hundreds of thousands of years.
This means that atmospheric carbon dioxide will be elevated by
human activities for a very long time – even if we stop burning
fossil fuels in the next few decades.
Methane is another important greenhouse gas – each molecule of
methane has the warming power of approximately twenty carbon
dioxide molecules. In the last decade, about half of the methane
emissions are due to human activities. These human emissions
have increased atmospheric methane from approximately 0.8 ppm
in 1750, before the industrial revolution, to 1.83 ppm in 2014.
Methane’s lifetime in the atmosphere is ten years, which is much
shorter than carbon dioxide.
188

Additional reading
Ciais et al., “Carbon and Other Biogeochemical Cycles,” in T. F. Stocker,
D. Qin, G.-K. Plattner, M. Tignor, S. K. Allen, J. Boschung, A. Nauels, Y.
Xia, V. Bex, and P. M. Midgley (eds.), Climate Change 2013: The
Physical Science Basis. Contribution of Working Group I to the Fifth
Assessment Report of the Intergovernmental Panel on Climate Change
(Cambridge and New York: Cambridge University Press, 2013). This is
the carbon cycle chapter from the latest IPCC report. As with most IPCC
reports, it is not the easiest thing to read, but if you want to know what
scientists think, this is where you should go.
D. Archer, The Global Carbon Cycle (Princeton, NJ: Princeton University
Press, 2010). This is a short and focused textbook on carbon cycle science.
It is a great introduction to the subject.
E. Roston, The Carbon Age: How Life’s Core Element Has Become
Civilization’s Greatest Threat (New York: Walker, 2009). This is a fun
and easy-to-read book about carbon and the immense role it plays in our
lives.
See www.andrewdessler.com/chapter5 for additional resources for
this chapter.
189

http://www.andrewdessler.com/chapter5

Terms
Carbon cycle
Carbonic acid
Chemical weathering
Deep ocean
Deforestation
Fossil fuels
Greenhouse gas
Halocarbons
Isotopes
Keeling curve
Land-use changes
Mixed layer
Ozone
Parts per million
Photosynthesis
Radiocarbon dating
Radiocarbon dead
Respiration
Time scale
190

Turnover time
191

Problems
1.
a) Describe the processes that transfer carbon from the atmosphere
to the land and from the land to the atmosphere. What are the
chemical reactions that describe these processes?
b) How do these processes interact to produce the “sawtooth”
annual cycle in the atmospheric abundance of CO2 shown in
Figures 5.1 and 5.6?
2. A letter to the editor of the Austin American-Statesman, published
on December 23, 2009, asks this question: “The trillion-dollar
question that Copenhagen has not answered [is this]: Because carbon
dioxide molecules are all identical, why is it that carbon dioxide from
carbonated beverages, pets, cattle, farm animals, and humans, yeast,
dry ice, fireplaces, charcoal grills, campfires, wildfires, alcohol and
ethanol is good, and carbon dioxide from fossil fuel is bad? Can
anyone in the United States answer this question?” What is your
answer?
3. Your aunt asks you how we know that volcanoes are not
responsible for the observed increase in carbon dioxide. What do you
tell her?
4. Explain how isotopes help us identify human activities as the
reason atmospheric carbon dioxide is increasing.
5. Your grandfather asks you to explain how humans are modifying
the carbon cycle. What do you tell him?
192

6. Explain how “chemical weathering” removes CO2 from the
atmosphere. What is the weathering chemical reaction? Can this
process play an important role in counteracting the increase in
atmospheric carbon dioxide caused by humans?
7. Of the carbon dioxide humans add to the climate, approximately
half is removed within a few decades. Where does it go? How would
it affect the climate if, all of the sudden, all of the carbon dioxide we
emit stayed in the atmosphere?
8. Why is rain naturally acidic? What then, does the term acid rain
refer to? (Acid rain is not covered in the chapter, so you will have to
do some outside research on it.)
9. Imagine that 100 molecules of carbon dioxide are produced by
combustion of fossil fuels; half of these molecules are immediately
absorbed by the land biosphere and ocean. A fraction f of this goes
into the land biosphere (so 1-f goes into the ocean). a) Make a plot of
the change in oxygen as f varies from 0 to 1. b) If the observed
change in oxygen is a decrease of 110 molecules, what fraction of
carbon was absorbed by the ocean? Assume that fossil fuel
combustion consumes 1.4 molecules of O2 for every molecule of CO2
produces; and photosynthesis produces 1.1 molecules of O2 for every
molecule of CO2 consumed; absorption of CO2 by the ocean has no
effect on O2.
10. The sawtooth in the CO2 time series due to the annual cycle in
northern hemisphere plant growth is dramatic. The data in Figure 5.1
comes from Mauna Loa, in Hawaii. Based on the material in this
chapter, predict how the magnitude of this annual cycle in the Arctic
would compare to that in the Antarctic. Find the data online
193

(http://andrewdessler.com/chapter5) and see if you can confirm your
hypothesis.
11. Why do the atmospheric oxygen measurements disprove an
oceanic or volcanic source for the increase in atmospheric carbon
dioxide?
1 Throughout this section, the percentages given are of volume, not
mass. For the chemists reading this, this is the same as mole fraction.
2 Parts per million can be by volume (number of molecules out of every
million) or by mass (grams of constituent out of every million grams of
air). Following the previous discussion, all mixing ratios in this book
will be by volume.
3 This means that, of every billion molecules of air, a few are ozone.
4 Cement production is lumped into this number; it constitutes a few
percent of this total.
5 By the way, there is so much oxygen in the atmosphere that this small
observed decrease should not worry you.
194

http://andrewdessler.com/chapter5

6
Forcing, feedbacks, and climate
sensitivity
◈
In Chapter 4, we showed that the temperature of a planet is a function of
the solar constant, the albedo of the planet, and the composition of the
atmosphere (Equation 4.5). In Chapter 5, we showed that humans are
adding greenhouse gases to the atmosphere, so we would expect the
planet’s temperature to be increasing. In Chapter 2, we showed that
temperature is indeed going up. If that were all there was to climate
change, we would be done with the science. But there is a lot more
interesting physics that we have to consider to fully understand modern
climate change.
195

6.1 Time lags in the climate system
In our climate calculations in Chapter 4, we discussed equilibrium
situations in which we explicitly assume that Ein and Eout are equal. But
modern climate change is not an equilibrium problem. To understand the
time-dependent behavior of the climate system, consider a planet with no
atmosphere that is in equilibrium (energy in equals energy out). Assuming
values for the Earth (Ein = 238 W/m2), we calculated in Chapter 4 that the
planet’s surface temperature would be 255 K. The energy fluxes for this
planet are diagramed in Figure 6.1a.
Figure 6.1 Schematic of energy fluxes on a planet (a) with no
atmosphere, (b) the instant after a one-layer atmosphere is added to the
planet, and (c) after the climate reaches its new equilibrium.
Now let us imagine that a one-layer atmosphere is instantly added to
the planet. What are the fluxes the instant after the layer is added? The
temperature of the surface is still 255 K because objects have thermal
inertia, which prevents their temperatures from changing instantly – just
like the inertia that keeps a car from stopping instantly when you hit the
brakes. Anyone who has worked in a kitchen knows that turkeys placed in
an oven do not cook instantly but take hours to heat up.
This means that, the instant after the atmosphere is added, the surface
is emitting exactly the same as it was before the atmosphere was added,
196

238 W/m2. The atmosphere, however, is now absorbing all of the photons
coming from the surface. Half of the absorbed energy is reemitted upward
to space, and half is reemitted downward back to the surface.1 This is
shown in Figure 6.1b.
The addition of the atmosphere has therefore reduced energy out for
the planet by half, to 119 W/m2. Given that energy in remains the same,
238 W/m2, energy in exceeds energy out for the planet, so the planet must
warm in response. Moreover, the energy that is no longer escaping to
space, 119 W/m2, has been redirected by the atmosphere back toward the
surface. As a result, energy in for the surface is now Ein = 238 + 119 = 357
W/m2, which exceeds energy out of Eout = 238 W/m2, so the surface is also
warming. As the surface and rest of the planet warms, energy out
increases. Eventually, the surface and atmosphere warm enough that Ein =
Eout and the planet is again in energy balance; this is shown in Figure 6.1c.
The scenario just described is basically what occurs when greenhouse
gases are added to our atmosphere. The greenhouse gases intercept some
of the energy escaping to space and redirect it back toward the surface. In
doing so, greenhouse gases both reduce energy out for the planet and
increase energy in for the surface, thereby knocking the system out of
equilibrium and forcing the climate system to warm. The planet warms
until energy out again balances with energy in for the planet as a whole
and for each individual component of the climate system.
How long does it take for the planet to reach a new equilibrium
temperature after the addition of greenhouse gases to the atmosphere?
Most of the thermal inertia of the climate system on Earth comes from the
ocean: it covers 70 percent of the Earth to an average depth of 4,000 m
(2.5 miles). However, you shouldn’t think of the ocean as a single entity.
Just like in Chapter 5, it is more accurate to consider the ocean split into
two parts: a mixed layer, a few hundred meters thick, that communicates
197

rapidly with the atmosphere and the deep ocean containing the other
roughly 95 percent of the ocean’s mass, which communicates slowly with
the atmosphere.
The mixed layer is in contact with the atmosphere, so greenhouse
warming directly heats it. And because its mass is relatively small
(compared to the mass of the entire ocean), it warms rapidly and reaches
equilibrium after just a few decades. After that, the rate of warming of the
climate system is controlled by the rate of heating of the deep ocean. This
deep-ocean warming is so slow that it takes millennia for the entire ocean
to reach equilibrium.
Thus, after release of a slug of carbon dioxide, you get rapid warming
for a few decades (as the mixed layer warms) followed by slower warming
for thousands of years (as the deep ocean warms). Because of this, the
carbon dioxide we are emitting today will still be warming the climate in
the year 3000 and beyond. A sobering thought, indeed.
This lag between emission of greenhouse gases and the resultant
warming has some important implications. Because we are constantly
emitting greenhouse gases, climate change is presently lagging the amount
of greenhouse gases in our atmosphere. So if we stopped emitting
greenhouse gases today, the climate would continue to warm for centuries.
Future warming from emissions that have already occurred is likely to be
about half a degree Celsius,2 comparable to the warming of the last
century. This is often referred to as committed warming or, more
informally, as warming “in the pipeline.” We have already paid for this
warming with emissions over the past few decades, but the warming has
not yet arrived. Nevertheless, it is coming – and there is little we can do to
avoid it.
The lag between emissions and warming also has important
implications for the policy debate over climate change. Because it takes
198

decades to experience the bulk of the warming from today’s emissions, the
benefits from reductions in emissions today, which may be costly, will
also not be fully realized for many decades. This means that reducing
emissions requires today’s society to pay costs that may primarily benefit
future generations. This creates an incentive for policymakers to do
nothing about climate change – “kick the can down the road,” as they say
– and leave the problem for people in the future to solve. When we explore
policy options later in the book, we will return to this problem.
199

(6.1)
6.2 Radiative forcing
Figure 6.2 plots energy out to space for the planet shown in Figure 6.1.
The period before Year 0 corresponds to Figure 6.1a, when the planet had
no atmosphere and it was in energy balance. At Year 0, the atmosphere is
instantaneously added, and energy out drops immediately to 119 W/m2.
This corresponds to Figure 6.1b. As the surface heats up in response to the
warming from the atmosphere, energy out increases, eventually reaching
238 W/m2 – and the planet is once again in energy balance, corresponding
to Figure 6.1c.
Figure 6.2 Plot of energy out for the planet shown in Figure 6.1. The
atmosphere is added instantaneously in Year 0.
This leads us to one of the most important concepts in climate
science: radiative forcing. Radiative forcing (often abbreviated RF) is the
change in Ein − Eout for the planet as a result of some change imposed on
the planet before the temperature of the planet has adjusted in response3:
In the example just given, ΔEout is −120 W/m2; that is the drop in Eout
the instant after the atmosphere is added but before the warming
200

temperature has caused Eout to increase. Note that ΔEin is zero because
energy in does not change when an atmospheric layer is added. Thus, the
radiative forcing of adding a one-layer atmosphere is 0 − (−120) = +120
W/m2. The sign convention is that positive radiative forcings correspond
to changes that warm the climate, whereas negative ones correspond to
changes that cool the climate.
An example: What is the radiative forcing of a 5 percent increase
in solar constant for the Earth that occurs over 100 years?
Let us begin by calculating ΔEin, the change in energy in. From
Chapter 4, we know that Ein = S(1 − α)/4, which for the Earth is
238 W/m2. If the solar constant S increased by 5 percent, then S
would increase to 1,360(1.05) = 1428 W/m2. For this new value of
the solar constant, Ein = 250 W/m2. Thus, Ein has increased from
238 W/m2 to 250 W/m2, so ΔEin = +12 W/m2.
What is ΔEout for this solar constant change? Eout is
determined entirely by atmospheric composition (i.e., number of
layers) and temperature. Atmospheric composition is not changing
in this example, and radiative forcing is defined as the response to
an instantaneous change, before the temperature of the planet has
adjusted to the change. Thus, Eout does not change and ΔEout = 0.
Putting it together using Equation 6.2, we see that the radiative
forcing for this change in the solar constant is +12 W/m2.
The fact that the change occurred over the course of 100 years
does not enter into the calculation. Radiative forcing calculations
are done under the assumption that the climate is not allowed to
respond to the change, so the length of time the change is imposed
over is irrelevant.
201

The imposition of a radiative forcing on a planet, such as a change in
solar constant, will take the planet out of energy balance – so that Ein and
Eout are no longer equal to each other. In response, the temperature of the
planet will adjust so that Ein once again equals Eout. In the case of an
increasing solar constant, the planet will warm, increasing Eout, until Ein
and Eout are once again in balance.
Thus, radiative forcing is a quantitative measure of how much some
climate perturbation (e.g., an increase in solar constant, increase in
greenhouse gases) will change the climate. The advantage of using
radiative forcing is that it allows us to express diverse changes to the
climate system by using a common metric. For example, it allows us to
compare the climate-changing effect of a 100-ppm increase in carbon
dioxide to a 1 percent increase in the solar constant. By comparing the
radiative forcing of these two changes, we could determine which one
would warm the planet more. Similarly, radiative forcing of +1 W/m2 will
produce similar warming of the climate, regardless of whether that change
was caused by a brightening of the Sun, an increase in carbon dioxide, an
increase in methane, or some other change.
Figure 6.3 shows the radiative forcing of the various factors that have
influenced our climate over the past few centuries. In much the same way
that temperature anomalies are the change in temperature from a reference
period, radiative forcings are generally calculated as a change from a
reference climate. In Figure 6.3, the values plotted are radiative forcing
caused by changes since 1750, which is considered the preindustrial value.
In the rest of this section, I describe each one of these factors.
202

Figure 6.3 Radiative forcing caused by human activities between 1750
and 2010. The error bars indicate the uncertainty of the estimate (based
on values from Ciais et al., 2013).
203

(6.2)
6.2.1 Greenhouse gases
The atmospheric abundance of carbon dioxide increased from 280 to 391
ppm between 1750 and 2010; this change reduces energy out by 1.82
W/m2 (energy in does not significantly change). Thus, the radiative forcing
of carbon dioxide is +1.82 W/m2. Increases in methane, nitrous oxide, and
the halocarbons between 1750 and 2010 produced radiative forcings of
+0.48, +0.16, and +0.36 W/m2, respectively, for a total of +1.00 W/m2.
Ozone in the lower atmosphere is both a greenhouse gas and one of
the primary components of photochemical smog. As the world has become
more industrialized, lower atmospheric ozone has increased along with the
other components of air pollution. This increase contributes a positive
radiative forcing of +0.4 W/m2. Ozone in the stratosphere, in contrast, has
been declining as a result of ozone depletion from halocarbons and other
manmade chemicals. This contributes a negative radiative forcing of −0.05
W/m2.
Next is stratospheric water vapor. An important source of
stratospheric water vapor is the transport of methane into the stratosphere
followed by oxidation, which has this net reaction:
The increase in methane over the past two centuries has therefore
increased stratospheric water, which has led to a positive radiative forcing
of +0.07 W/m2. We will consider lower-atmospheric water vapor in the
section on feedbacks.
Thus, although carbon dioxide is the single most important
greenhouse gas emitted by human activities, it is not the only important
one. In fact, the combined radiative forcing from the other greenhouse
204

gases (+1.42 W/m2) is nearly as large as the radiative forcing from carbon
dioxide alone (+1.82 W/m2). This has important implications for policies
to address climate change, as we will discuss later.
205

6.2.2 Aerosols
Aerosols are particles so small that they do not fall under the force of
gravity but remain suspended in the atmosphere for days or weeks.
Aerosols can interact both with sunlight that is falling on the planet and
with infrared radiation that is being emitted by the surface and atmosphere
– thereby altering the climate. There are several types of aerosols, and their
composition determines how they interact with sunlight and infrared
radiation.
When fossil fuels containing sulfur impurities are burned, the sulfur is
released to the atmosphere with the other products of combustion. Sulfur is
also released into the atmosphere during biomass burning and from natural
processes in the ocean. Once in the atmosphere, the sulfur gases react with
other atmospheric constituents to form small liquid droplets, known as
sulfate aerosols.
Sulfate aerosols are highly reflective and reflect incoming solar
radiation back to space, so their net effect is to cool the climate. As a result
of increases in fossil fuel use over the past two centuries, the abundance of
sulfate aerosols has steadily increased with time, providing a negative
radiative forcing of −0.4 W/m2.
Such sulfate aerosols occur in the lower atmosphere, so their lifetime
is short – it takes just a few weeks before the aerosols are either washed
out of the atmosphere by rain or fall to the ground. This means that the
radiative forcing the Earth is experiencing at any given time from these
aerosols is due entirely to emissions of sulfur that occurred in the past
month or two.
Another important – but episodic – source of sulfur gases for the
atmosphere is volcanic eruptions. Volcanoes emit enormous amounts of
206

sulfur gas, and energetic eruptions can inject it directly into the
stratosphere. Aerosols in the stratosphere can remain there for several
years – much longer than an aerosol resides in the lower atmosphere. This
long lifetime, combined with the massive amounts of sulfur released,
means that a single volcano can produce a negative radiative forcing of
several watts per square meter that lasts for several years after the eruption
(Figure 6.4).
Figure 6.4 Radiative forcing from volcanoes
(obtained from NASA GISS; see
data.giss.nasa.gov/modelforce/Fe.1880–2011.txt).
This negative radiative forcing can lead to a noticeable cooling of the
climate following an eruption. In 1816, for example, after three major
eruptions in three years, the United States and Europe experienced the
“year without a summer,” in which snow fell in Vermont in June and
heavy summer frosts caused crop failures and widespread food shortages.
When that summer was followed by a winter so cold that the mercury in
thermometers froze (this happens at −40°C), many residents fled the
Northeast United States and moved south.
A few years after a volcanic eruption, volcanic aerosols fall out of the
stratosphere and the climate warms back up. Combined with the fact that
207

such massive volcanic eruptions occur infrequently (as we can see in
Figure 6.4), the long-term impact of volcanoes on the climate has been
relatively small over the past few centuries.
Black carbon aerosols, such as soot, are another important aerosol
type. This type of aerosol is produced by incomplete combustion, such as a
smoldering fire or by two-stroke gasoline engines, so they are generally of
human origin. Because they are dark, they absorb solar radiation and
decrease the planet’s albedo, thereby warming the planet. Over the past
few centuries, black carbon aerosol abundance has increased as more
people burn more stuff, leading to a positive radiative forcing of +0.4
W/m2. Much like sulfate aerosols, these black carbon aerosols have
atmospheric lifetimes of a few weeks.
Another type of aerosol is mineral dust. Most of this dust comes from
natural processes, such as dust picked up off the world’s deserts by strong
winds (Figure 6.5). But approximately 20 percent of mineral dust comes
from anthropogenic sources – mainly agricultural practices (e.g.,
harvesting, plowing, overgrazing), changes in surface water features (e.g.,
drying out of lakes such as the Aral Sea and Owens Lake) and industrial
practices (e.g., cement production, transport). The net effect of dust is to
cool the planet, and human activities have contributed a negative radiative
forcing of −0.1 W/m2. Like other types of aerosols, these dust aerosols
have atmospheric lifetimes of a few weeks.
208

Figure 6.5 Image of a strong temperate cyclone over China, pushing a
wall of dust as it moved. The image was captured in early April of 2001
by the Moderate Resolution Imaging Spectroradiometer on NASA’s
Terra satellite (image obtained from the Earth Observatory; see
earthobservatory.nasa.gov/IOTD/view.php?id=8341).
Combining all types of aerosols (those discussed earlier and several
not discussed), the direct radiative effect of aerosols is to cool the climate,
with an estimated negative radiative forcing of −0.35 W/m2. However,
aerosols also have an indirect effect on the climate, whereby aerosols
influence the climate by altering clouds. There are several ways that this
can occur. One of the clearest mechanisms is by altering the number of
particles in a cloud. Cloud particles generally form when water condenses
onto cloud-condensation nuclei or CCN, which are small solid or liquid
aerosols that are hydrophilic, meaning that they attract water. It is the
number of CCN in an air mass that determines how many cloud particles
are found in a cloud.
209

If you add aerosols to a cloud, then you will increase the number of
CCN – and therefore the number of droplets making up the cloud. But the
total liquid water contained in the cloud is fixed. Thus, the increase in the
number of droplets means that each droplet has less water and is therefore
smaller. This is akin to cutting a pie into more slices: The total amount of
pie is fixed, so more slices means that each slice must be smaller.
It turns out that a cloud containing smaller droplets is more reflective
than one containing large droplets. A familiar example of this can be seen
in your kitchen in the difference between regular table sugar and powdered
sugar. Chemically, the two substances are identical, but powdered sugar is
made up of much smaller particles. Because smaller particles tend to be
more reflective, the pile of powdered sugar is a brighter white than a pile
of table sugar.
Thus, if one adds aerosols to a cloud, the cloud gets brighter and more
reflective. This can be seen in what are called ship tracks. The exhaust
from diesel engines contains fine particulates that can serve as CCN. As
ships steam across the ocean, these fine aerosol particles from their
engines are transported by the winds into low-level clouds, leading to
increases in numbers of droplets and brighter clouds. From a satellite
(Figure 6.6), lines of bright clouds trace out the paths of these ships.
210

Figure 6.6 Ship tracks in clouds off of the West Coast of the United
States (image obtained from the Earth Observatory, see
earthobservatory.nasa.gov/IOTD/view.php?id=37455).
This effect on cloud reflectivity is just one effect aerosols have on
clouds. By making the cloud particles smaller, aerosols also slow down the
coagulation process whereby cloud droplets combine to form raindrops.
This reduces rainfall from a cloud, so the clouds last longer. Aerosols can
also change the height of the cloud, as well as the phase (ice vs. liquid).
The addition of black carbon to clouds can lead to local warming that can
cause clouds to evaporate.
Considering all the effects of aerosols on clouds, scientists estimate
that the indirect aerosol effect produces a negative radiative forcing of
−0.45 W/m2. However, as the error bars in Figure 6.3 show, the indirect
effect of aerosols is highly uncertain. Combining the direct and indirect
effects, aerosols produce a negative radiative forcing of roughly −0.9
W/m2.
211

Because aerosols last only a few weeks in the atmosphere before they
are removed, aerosols do not have time to become well mixed throughout
the atmosphere (which takes a year or so). As a result, the distribution of
aerosols in our atmosphere is highly variable, with most aerosols found
near their sources. Figure 6.7 shows their distribution, and from this we
can infer the major sources of aerosols across the globe: Saharan dust that
is blown westward from North Africa into the Atlantic, smoke from
biomass burning over central Africa, and an aerosol soup that originates
over Asia and is blown eastward over the Pacific toward the United States.
Figure 6.7 Annual average aerosol optical depth (a measure of the
abundance of aerosols) as a function of latitude and longitude, for the
years 2004–2008 (measurements were made by the Moderate
Resolution Imaging Spectroradiometer onboard NASA’s Aqua satellite
and were obtained from the NASA Goddard Earth Sciences Data and
Information Services Center). White areas are regions where no data
were obtained.
It is also apparent that aggressive air-pollution control efforts in the
United States and Western Europe have worked – these regions have low
levels of aerosol abundance despite large economies. It is the poorer
countries with lax environmental regulations, such as China and India, that
are responsible for much of the anthropogenic aerosols in the atmosphere.
212

So although the global average radiative forcing from aerosols is −0.9
W/m2, this is not evenly distributed over the globe. In regions where
aerosol abundance is high, aerosols can have a local radiative forcing of
many times this value, whereas in regions that have no aerosols, the local
radiative forcing can be zero. This can be contrasted to greenhouse gases
such as carbon dioxide or methane, which are well mixed in the
atmosphere because of their long atmospheric lifetimes (many years),
resulting in a radiative forcing evenly distributed across the globe.
From a climate perspective, the negative radiative forcing from
aerosols offsets 25 percent of the positive radiative forcing from
anthropogenic greenhouse gases. In this way, aerosols benefit us, because
without them the net radiative forcing would be higher – and global
warming would be worse. But aerosols are not all good – they are also one
of the main components of air pollution around the world, which kills
millions of people every year.
As these poorer countries begin to clean up their atmosphere, we
expect to see the amount of aerosols in the atmosphere diminish. Although
such reductions improve public health, they also make global warming
worse by reducing the cooling that aerosols provide. This is another factor
that must be considered in climate change policy.
213

6.2.3 Land-use changes
Humans have been remaking the face of the planet for thousands of years,
and over the past few centuries, humans have altered vast swaths of the
surface. For example, in 1750, approximately 7 percent of the global land
area was under cultivation or pasture; in 1990, that number was slightly
more than a third. Such alterations of the surface can modify the Earth’s
climate. Agricultural land typically has a higher albedo than does the
natural landscape, especially if the latter is forest. Thus, cutting down a
forest and replacing it with grassland for grazing cattle will increase the
surface’s albedo. Over all, human land-use changes have tended to cool the
planet, with a radiative forcing of −0.15 W/m2. This albedo effect is in
addition to any radiative forcing from carbon emissions from land-use
changes, which were discussed in Section 5.4.
Another way human activities can alter the albedo of the surface is
through the release of black carbon – basically soot. In the previous
section we explored how these dark particles have a warming effect when
suspended in the atmosphere because they absorb sunlight that falls on
them. However, their climate impact does not end there. After a month or
two in the atmosphere, these particles are removed from the atmosphere
and deposited on the surface. If the surface is bright, like snow, then this
deposition will reduce the albedo of the surface, thereby increasing the
absorption of solar energy. With the increase in industrial activities over
the past two centuries, the amount of black carbon deposited on snow has
led to a positive radiative forcing with an estimated magnitude of +0.04
W/m2.
214

6.2.4 Changes in the Sun
A final radiative forcing – and one that is entirely unrelated to humans –
comes from changes in the amount of solar energy reaching the Earth. As
we will explore in more detail in Chapter 7, rather than being a constant
source of energy, the Sun has well-documented variations. The best known
is the eleven-year sunspot cycle, over which the solar constant varies by
roughly 1 W/m2 (out of 1,360 W/m2, so it is not a terribly big variation).
The Sun’s output also varies slightly with a period of twenty-seven days,
which is the time it takes the Sun to rotate once on its axis. Because of the
large heat capacity of the climate system, neither of these variations in the
Sun have much effect on the climate.
However, it is possible that, over longer periods, the Sun’s output can
vary – and this might have an important influence on climate.
Unfortunately, our knowledge of these longer-term variations is poor.
Accurately measuring the output of the Sun must be done from orbit, and
that means that we do not have reliable measurements of the solar constant
before the late 1970s, when the first satellites designed to measure the
Sun’s output were launched. Scientists have, however, attempted to
reconstruct the long-term record of the solar constant by using proxy data,
such as sunspot number. This work has suggested that the Sun may have
gotten slightly brighter over the past 250 years, particularly early in the
twentieth century, and provided a positive radiative forcing estimated to be
+0.05 W/m2.
215

6.2.5 Total net forcing
Summing all of the radiative forcings in the period between 1750 and
2005, we get a net radiative forcing of +2.3 W/m2. Considering
uncertainties (expressed as error bars in Figure 6.3), the total radiative
forcing could actually be anywhere between +1.1 and +3.3 W/m2. This
total comes from positive forcings, primarily the greenhouse gases, and
negative forcings from aerosols and land-use changes.
Let me emphasize what this value means: If we made all of the
changes in greenhouse gases, aerosols, land use, and so on instantaneously
and without letting the atmosphere warm, Ein − Eout would have increased
2.3 W/m2. In order for the Earth to once again be in energy balance, the
planet would have to warm, which would increase Eout. Earlier in the
chapter, I said that, for the present-day Earth, Ein exceeds Eout by
approximately 0.5 W/m2. This means that the planet has warmed up
enough in the past 250 years to erase +1.8 W/m2 of the radiative forcing.
The planet needs to continue warming in order to erase the remaining
radiative forcing – this is warming that we are already committed to and
can do little to stop.
Another important take-away message from this section is that
climate change is about much more than carbon dioxide. Carbon dioxide is
indeed the most important greenhouse gas that humans are adding to the
atmosphere and the most important climate forcing of the past few
centuries. But other greenhouse gases that humans are adding are also
making important contributions to the radiative forcing. And non-gas
constituents, such as aerosols, also play an important role.
You should now be able to see the foundation of the policies we
might undertake to stabilize our climate. The temperature of the climate is
216

on an upward trajectory because we are increasing the net radiative forcing
through emissions of greenhouse gases. If we stabilize net radiative
forcing, then the climate will stabilize a few decades later. There are two
obvious ways to do this. First, we can stop activities that produce positive
radiative forcing – this basically means ceasing emissions of greenhouse
gases to the atmosphere. In the policy world, this is known as mitigation.
Alternatively, we could intentionally engage in activities that produce
negative radiative forcing – for example, we could add sulfate aerosols to
the atmosphere, thereby canceling out the positive radiative forcing. This
latter approach is what is commonly referred to as geoengineering. I will
have more to say about both approaches in Chapters 11 and 12.
217

6.3 Feedbacks
One of the ultimate goals of climate science is to make quantitative
predictions of the future climate. As discussed in Chapter 4, given the solar
constant, albedo, and atmospheric composition, it is conceptually easy to
calculate the temperature of a planet. You may expect, therefore, that it
would be pretty easy to calculate how much warming we should expect if
we applied some specified radiative forcing to the planet (e.g., carbon
dioxide is doubled, the Sun gets 1 percent brighter).
For example, consider a hypothetical planet with no atmosphere (n =
0) and an Earth-like solar constant and albedo (S = 1,360 W/m2, α = 0.3).
The surface temperature of this planet is, according to Equation 4.5, T =
254.5 K. Now imagine that greenhouse gases are added to the atmosphere,
so that n increases to 0.01. According to Equation 4.5, the new temperature
would be 255.2 K, which is a warming of 0.6 K.
That answer would be correct if nothing else in the climate system
changed. But as the planet warms, other things do change. For example,
because ice melts reliably at 0°C, global warming should reduce the
amount of ice on the Earth’s surface – and, as discussed in Section 2.1.3,
that is indeed happening. If the melting ice uncovers a dark surface, such
as ocean, then this decreases the average planetary albedo (i.e., makes the
planet less reflective), which leads to more absorption of solar radiation
and additional warming. This additional warming leads to even more
melting, which leads to further decreases in albedo and further warming,
and so on. This is known as a feedback loop, and it is shown schematically
in Figure 6.8.
218

(6.3)
Figure 6.8 The ice-albedo feedback loop.
The net effect of the feedback loop shown in Figure 6.8 is to amplify
the initial warming from the addition of greenhouse gas. To provide a
quantitative demonstration of this, let us make the reasonable assumption
that the albedo is proportional to the planet’s surface temperature and that
the relation between these two variables is: α(T) = 0.3 − (T − 254.54)⁄150.
This relation quantifies the change in albedo as the planet’s temperature
changes.
The equation for surface temperature now includes an albedo that is a
function of T:
This is no longer a trivial equation to solve because T
appears on both sides of the equation. Rather, this must be solved by
factoring a fourth-order polynomial or by solving the equation
numerically.
Using one of those techniques, we can calculate the temperature of a
planet with n = 0 and S = 1,360 W/m2 (we no longer specify albedo
because it is a function of temperature). Solving this equation, we find that
the temperature is 254.5 K, the same as when we assumed the albedo α =
0.3. This is not an accident because we defined the albedo to have a value
of 0.3 at that temperature.
219

Now let us add the same amount of greenhouse gases to the
atmosphere as we did in the no-feedback example, which increases n to
0.01. Now the solution to Equation 6.4 is T = 256.2 K, which yields a
warming of 1.7 K. In other words, the inclusion of melting ice raises the
warming from 0.6 K to 1.6 K. This amplification of the warming is
referred to as a positive feedback.
You can also have a negative feedback, which acts to reduce the
initial warming. Imagine a planet that is covered with flowers of two
colors: white and black. As the temperature of the planet goes up, the
white flowers prosper while black flowers die. This means that, as a planet
warms, the planet also becomes whiter – i.e., the albedo goes up.
To quantitatively explore the effects of this, let us assume that the
relation between albedo and temperature on this flower-covered planet is
α(T) = 0.3 + (T − 254.5)⁄150, so that an increase in T leads to an increase in
α. In this case, an increase in greenhouse gases from n = 0.0 to n = 0.01
leads to a new surface temperature of 254.9 K, which is a warming of only
0.4 K. This is less warming than we would get if the albedo were constant.
Thus, the flowers have reduced the amount of warming and this flower
feedback is negative.
220

6.3.1 Fast feedbacks
There are several important feedbacks in our climate system. The feedback
just described, in which warmer temperatures melt ice, leading to reduced
albedo and additional warming, is known as the ice-albedo feedback.
Another important feedback is the water-vapor feedback, which arises
because a warmer atmosphere can hold more water vapor. Thus, global
warming leads to increased atmospheric humidity, and because water
vapor is itself a greenhouse gas, this leads to additional warming. Both the
ice-albedo and water-vapor feedbacks are positive, meaning that they
amplify an initial warming.
There are also negative feedbacks in the climate system. The biggest
negative feedback is known as the lapse-rate feedback. Because power
radiated by a blackbody is equal to σT4, a warmer atmosphere radiates
more power to space. Therefore, as the upper atmosphere warms, the
enhanced radiation offsets some of the initial warming.
The biggest debate among scientists today is about cloud feedback.
Clouds affect the climate in two opposite ways. First, they reflect sunlight
back to space, reducing energy in, which tends to cool the climate. Second,
they absorb infrared radiation emitted by the surface, decreasing energy
out just like a greenhouse gas, and this tends to warm the climate. The net
effect of clouds on the climate is therefore the difference of these two
opposing effects. In our present climate, the reflection of solar radiation is
slightly larger than the heat trapping effects, so clouds reduce Ein − Eout for
the Earth by roughly 25 W/m2.
This could, however, change in the future. If, in response to an initial
warming, the cooling effect of clouds is enhanced, then the effect of clouds
on the planet’s energy budget will be more negative and clouds will act to
221

reduce the initial warming and therefore be a negative feedback. In
contrast, if the heat trapping is enhanced, then the cloud’s radiative effect
will become less negative and clouds will amplify the initial warming and
therefore be a positive feedback. Although the exact answer is uncertain,
the best guess of the scientific community is that the climate feedback is
positive, meaning that clouds amplify warming.
Feedbacks discussed in this section are known as fast feedbacks
because they occur rapidly enough in response to a change in surface
temperature that they will play an important role in the evolution of
climate change over the coming century. Water vapor, clouds, and the
lapse rate all respond within a week or so to changes in surface
temperature, and therefore their impact on energy in and energy out is fast.
The response time of ice depends on the type. Some types (e.g., snow) will
respond in weeks or months while others (e.g., sea ice) will respond in
years, so changes in these types could also be considered a fast feedback.
222

6.3.2 Slow feedbacks
In contrast to fast feedbacks, slow feedbacks include processes that
respond slowly to increasing surface temperature, so they require long
periods of warmth before they significantly alter energy in or energy out.
For example, the Greenland and Antarctic ice sheets are so big that they
will likely take many centuries to significantly respond to a change in
temperature. The ice-albedo feedback associated with ice sheets would
therefore be categorized as a slow feedback.
Another slow feedback revolves around the fact that there are large
amounts of carbon stored in the ground. One of these carbon reservoirs is
permafrost, which was discussed in Section 5.2.1. This permafrost
contains dead organic plant matter that is kept intact as long as the ground
remains frozen. If a warming climate leads to the melting of permafrost,
then the organic matter in it thaws out and decays, releasing the carbon
back into the atmosphere in the form of either methane or carbon dioxide.
We know that permafrost is indeed melting, which is consistent with
the large warming in the Arctic over the past few decades (e.g., see Figure
2.2b). In Alaska, for example, roads and buildings constructed on
permafrost under the assumption that the permafrost would never melt are
now suffering damage as the permafrost melts and the ground underneath
begins shifting. In Siberia, melting permafrost formed during the last ice
age is revealing frozen oddities such as intact woolly mammoths that died
20,000 years ago.
Another source of frozen greenhouse gases is what are known as
methane clathrates – methane molecules that are embedded in ice.
Clathrates can exist on land or under the ocean, and as with permafrost,
warming temperatures can melt the ice and release the methane trapped
223

therein. Given that a molecule of methane is twenty times as powerful a
greenhouse gas as a molecule of carbon dioxide, the rapid release of
significant amounts of methane is a worrying possibility. And there are
several other reservoirs of carbon, such as the tropical forests, which may
release the carbon to the atmosphere as the climate warms.
This opens the possibility of a carbon-cycle feedback, in which an
initial warming leads to the release of large amounts of carbon dioxide and
methane that are currently frozen in the ground or otherwise sequestered.
The release of these greenhouse gases leads to more warming, and the
further release of greenhouse gases. The occurrence of such a feedback in
the next few centuries is speculative, but there is reasonably strong
evidence that they have occurred in the past, such as during ice-age cycles.
I will return to this point when I discuss ice ages in Chapter 7.
Another slow feedback involves vegetation. It has long been known
that the distribution of vegetation on the Earth’s surface is governed to a
large extent by the climate – through the distribution of precipitation,
temperature, sunlight, and other such factors. Recently, however, it has
been realized that changes in vegetation can also affect the climate. For
example, the conversion of a forest to grassland will increase the albedo
(because the forest is darker than the grassland), thereby tending to cool
the climate. Changes in vegetation can also directly impact exchanges of
heat, water, and momentum between the surface and atmosphere, or
modify the rate of uptake of carbon dioxide by the vegetation. This
introduces the possibility of vegetation feedbacks in which changes in
climate lead to changes in vegetation, which in turn lead to additional
changes in climate.
Probably the slowest feedback is the weathering thermostat. As the
Earth’s surface warms, the total amount of rainfall will also increase. The
increase in rainfall in turn increases the rate of chemical weathering, which
224

removes carbon dioxide from the atmosphere (we explored this in Section
5.3). The reduction in atmospheric carbon dioxide acts to offset some of
the initial warming, so this is a negative feedback that tends to stabilize the
Earth’s temperature. That is the good news. The bad news is that the
weathering thermostat operates on geologic time scales, so it only has an
impact on the climate over millions of years. We should not expect the
weathering thermostat to ameliorate warming over the next century – or
over any time period that human society cares about.
In general, slow feedbacks are much more uncertain than fast
feedbacks because they are so slow that modern Earth science, which is
really only a few decades old, simply does not have data extending over
periods long enough to observe, understand, and quantify them. Thus, the
net effect of slow feedbacks on the climate is uncertain. Nevertheless, they
continue to compel our attention because many of the worst-case climate
scenarios involve slow feedbacks causing extremely large warming in the
next few centuries.
An aside: Feedback vs. radiative forcing
It is worth explicitly discussing the differences between climate
feedbacks and radiative forcings. Feedbacks are processes that
respond to changes in the Earth’s surface temperature, so feedbacks
do not initiate climate change. Rather, positive feedbacks amplify
and negative feedbacks ameliorate an initial warming. Water
vapor, for example, is considered a feedback because the amount
of water vapor in the atmosphere is set by the surface temperature
of the Earth. If the surface temperature increases, then the amount
of water in the atmosphere will also increase, leading to additional
warming.
225

Radiative forcings, in contrast, affect the climate but are
themselves unaffected by the climate. The changes in carbon
dioxide, methane, and the like between 1750 and 2005 are
fundamentally unrelated to the Earth’s temperature; instead they
are driven by economic activities of human society. Ditto for
aerosols.
A confusion arises because some things can be both
feedbacks and forcings. For example, although carbon dioxide has
been a forcing over the past two centuries, it can also be a feedback
if warming temperatures lead to the release of carbon dioxide.
Changes in vegetation are a forcing when humans are modifying
the vegetation, but they are a feedback when it is the climate that
causes the modification.
In most cases, it is clear whether a process is a forcing or
feedback. The increase in carbon dioxide over the last two
centuries is clearly driven by human activities, not surface
temperature, so it is certainly a forcing. Other processes are more
ambiguous. For example, the processes that regulate stratospheric
water vapor are not well understood. As a result, we do not know if
changes in stratospheric water vapor that are now occurring are
due directly to human activities, so they would be a forcing, or
indirectly, through changing surface temperature, in which case
they would be a feedback. In this chapter, we put it in the forcing
category, but more subsequent scientific research may reveal that it
belongs in the feedback category. Or perhaps it belongs in both
categories.
226

6.4 Climate sensitivity
227

(6.4)
(6.5)
(6.6)
6.4.1 Feedback math
To get more quantitative about feedbacks, let us first go over some basic
feedback math. Consider a feedback loop, such as the ice-albedo feedback
(Figure 6.8). We will express the strength of this feedback as g, which is
the additional fractional warming produced by one trip through the
feedback loop per degree of initial warming. Thus, in response to an initial
warming ΔTi, the first trip through the feedback loop produces additional
warming of gΔTi. But the feedback also operates on this additional
warming gΔTi, and this produces an additional warming of g(gΔTi) =
g2ΔTi. And feedbacks operate on this additional warming too, leading to
an additional warming of g3ΔTi, etc. This goes on forever, so the final
warming ΔTf is:
We can write this more compactly as
The math ninjas among you will recognize that this infinite series can be
rewritten more simply as
If g = 0, then there is no feedback and the final temperature change is
equal to the initial temperature change. If g is between 0 and 1, then ΔTf is
larger than ΔTi, meaning the net sum of feedbacks is positive. If g is less
than 0, then ΔTf is less than ΔTi, meaning the net sum of feedbacks is
negative.
228

(6.7)
As positive feedbacks get stronger and g approaches 1, the
denominator in Equation 6.7 approaches 0 and ΔTf approaches infinity.
This should make sense from visual inspection of Equation 6.5: If g = 1,
then each subsequent term is as big as the previous one and, because the
series is infinite, the sum must also be infinite. This situation is sometimes
referred to as a “runaway greenhouse effect.” In that case, an initial
temperature perturbation leads to a very, very large temperature rise.
In Equations 6.5–6.7, g is the sum of the feedback parameters from
the individual feedbacks:
where gia is the ice-albedo feedback, gwv is the water-vapor feedback,
gcloud is the cloud feedback, and glr is the lapse-rate feedback (we consider
here only the fast feedbacks).
The strongest feedback is the water-vapor feedback, with a magnitude
gwv = 0.6. This feedback is big enough that, by itself, it would more than
double the initial warming ΔTi. The ice-albedo feedback is substantially
weaker, with a magnitude gia= 0.1. Because it is a negative feedback, the
lapse-rate feedback has a negative magnitude glr = −0.3. Finally, the cloud
feedback is quite uncertain, but most scientists would put its magnitude
gcloud = 0.0–0.3.
Summing these individual feedbacks, we get a total feedback
parameter for our climate of g = 0.4–0.7. This means that, for our climate,
ΔTf = 1.6–3.3 ΔTi. Thus, as much as two-thirds of the warming we
experience comes from feedbacks rather than the direct heating from
greenhouse gases. This is why feedbacks occupy much of the scientific
debate over climate change.
229

6.4.2 Sensitivity
For historical reasons, climate sensitivity is most frequently expressed as
the warming that occurs if the carbon dioxide is instantaneously increased
from 280 ppm, the preindustrial value, to 560 ppm, twice the preindustrial
value, and then one lets the climate reach a new equilibrium.
For this doubling of carbon dioxide, the initial warming ΔTi is 1.2°C.
Using the feedback strengths from the previous section implies a range of
final temperature ΔTf = 2–4°C. A more sophisticated analysis by the IPCC
concludes that the climate sensitivity is likely in the range 1.5°C to 4.5°C.
Doubled carbon dioxide corresponds to a radiative forcing of roughly
+4 W/m2. Given this, we can also express the sensitivity as the warming
per unit of radiative forcing. The climate sensitivity in these units is
0.38–1.1°C/(W/m2). If you need a single value for the climate sensitivity, I
use 0.75°C/(W/m2), which is a convenient value that falls near the middle
of the range.
As an example, in Section 6.2 we calculated that the radiative forcing
for a 5 percent increase in solar constant is +12 W/m2. Given that, we can
calculate how much warming we get by multiplying this radiative forcing
by the climate sensitivity of 0.75°C/(W/m2), yielding a warming of 9°C.
These values of the climate sensitivity only include the fast
feedbacks. This is probably appropriate for climate change over the next
century. Over several centuries or longer, though, the contribution of slow
feedbacks can become important. These feedbacks are thought to be
mainly positive, so the climate sensitivity may be significantly higher
when we consider such longer periods. Exactly how much higher is
unknown, but looking at previous long-term warming events in response to
greenhouse gas emissions (such as the Paleocene-Eocene Thermal
230

Maximum, or PETM, mentioned in Section 2.2.2) provides some evidence
that slow feedbacks may increase the climate sensitivity. That would be
very bad news for our descendants living several hundred years from now.
231

6.5 Chapter summary
A radiative forcing is an imposed change on the energy balance of
the Earth. It is calculated as the change in energy balance for the
planet (energy in minus energy out) after the imposition of the
specific change in the climate but before the climate has changed in
response.
In response to a radiative forcing, the Earth’s temperature adjusts so
that energy balance is reestablished.
Because of the Earth’s thermal mass, this adjustment happens over
time. There is a relatively rapid warming during the first few
decades as the ocean’s mixed layer warms. After that, it is the
warming of the deep ocean that sets the pace of warming. Given
the enormous heat capacity of the deep ocean, this warming takes
millennia. Thus, the climate will still be warming in 1,000 years as
a result of greenhouse gases we emit today.
The increase in greenhouse gases since 1750 has imposed a
radiative forcing of +3.2 W/m2. The increase in carbon dioxide is
responsible for +1.8 W/m2, or more than half of the total forcing.
The change in aerosols since 1750 has imposed a net radiative
forcing of −0.9 W/m2. This means that aerosols offset
approximately 25 percent of the radiative forcing from increasing
greenhouse gases. Summing all changes, we get a net radiative
forcing over this time period of +2.3 W/m2.
Positive feedbacks amplify and negative feedbacks ameliorate an
initial warming. For the problem of modern climate change, we are
232

mainly concerned with the following fast feedbacks: water vapor,
ice-albedo, lapse rate, and clouds. Together, they double to triple
an initial warming.
Feedbacks are processes that respond to changes in the surface
temperature, whereas forcings are unrelated to the surface
temperature. Thus, feedbacks do not initiate climate change, but
forcings do.
The Earth’s climate sensitivity, which is conventionally defined as
the equilibrium temperature increase caused by a doubling of
carbon dioxide from 280 ppm to 560 ppm, is 1.5–4.5°C. In terms
of radiative forcing, the climate sensitivity is 0.38–1.1°C/(W/m2), a
range centered on 0.75°C/(W/m2).
233

Additional reading
Myhre, G., D. Shindell, F.-M. Bréon, W. Collins, J. Fuglestvedt, J. Huang,
D. Koch, J.-F. Lamarque, D. Lee, B. Mendoza, T. Nakajima, A. Robock,
G. Stephens, T. Takemura, and H. Zhang, “Anthropogenic and Natural
Radiative Forcing,” in T. F. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S.
K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex, and P. M. Midgley
(eds.), Climate Change 2013: The Physical Science Basis. Contribution of
Working Group I to the Fifth Assessment Report of the Intergovernmental
Panel on Climate Change (Cambridge and New York: Cambridge
University Press, 2013). This is the IPCC’s latest summary of what we
know about radiative forcing (download at www.ipcc.ch). It is long, and
not always easy to read, but it is comprehensive.
Boucher, O., D. Randall, P. Artaxo, C. Bretherton, G. Feingold, P. Forster,
V.-M. Kerminen, Y. Kondo, H. Liao, U. Lohmann, P. Rasch, S.K.
Satheesh, S. Sherwood, B. Stevens, and X.Y. Zhang, “Clouds and
Aerosols,” in T. F. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S. K. Allen,
J. Boschung, A. Nauels, Y. Xia, V. Bex, and P. M. Midgley (eds.), Climate
Change 2013: The Physical Science Basis. Contribution of Working
Group I to the Fifth Assessment Report of the Intergovernmental Panel on
Climate Change (Cambridge and New York: Cambridge University Press,
2013). This is the IPCC’s latest summary about feedbacks, clouds, and
aerosols (download at www.ipcc.ch). Among other things, it contains a
thorough summary of how clouds and aerosols interact.
See www.andrewdessler.com/chapter6 for additional resources for
this chapter and www.andrewdessler.com/computer for computer exercises
that illustrate some of the important concepts, including the ocean’s role in
regulating the pace of climate change.
234

http://www.ipcc.ch

http://www.ipcc.ch

http://www.andrewdessler.com/chapter6

http://www.andrewdessler.com/computer

Terms
Carbon-cycle feedback
Climate sensitivity
Cloud-condensation nuclei
Cloud feedback
Committed warming
Direct radiative effect of aerosols
Fast feedbacks
Feedback
Ice-albedo feedback
Indirect effect of aerosols
Lapse-rate feedback
Radiative forcing
Ship tracks
Slow feedbacks
Thermal inertia
Vegetation feedback
Water-vapor feedback
235

Problems
1. When considering how long it takes for radiative forcing to warm
up the planet, it is useful to think about the heat capacity of the
climate system, which tells us how many Joules are required to raise
the temperature by one degree Kelvin.
The heat capacity of the climate system comes mainly from the
ocean, and we can estimate it to be 6 × 1024 J/K (this means that, if
you add 6 × 1024 J to the Earth system, the climate will warm by 1
K).
a. Imagine you impose a radiative forcing of +2.3 W/m2, what rate
of warming will result? Express your answer in °C/century.
b. The actual rate of warming of the planet’s surface is predicted to
be faster than this over the next century. Why?
2. Imagine a planet where S = 1360 W/m2 and α = 0.3.
a. If n = 0, what is the temperature of this planet?
b. If S increases to 1370 W/m2 and n remains equal to 0, what is
the new temperature?
c. Let us include a water vapor feedback: S increases to 1370
W/m2, but the number of layers, n, is a function of surface T: n(T)
= (T − 254.5)⁄100. What is the new surface temperature?
d. Using the answers to b and c and assuming that no other
feedbacks are operating, what is the value of g for this climate?
3. List the important fast feedbacks operating in our climate. Identify
whether each is positive or negative.
236

4. Define climate sensitivity. What is the currently accepted value for
our climate?
5. Imagine that we add some carbon dioxide to the atmosphere and
the Earth warms by 1°C. How much warming would there have been
if there were no feedbacks?
6. Imagine that our Sun brightens by 1 percent instantaneously.
a) How long would it take for the Earth to reach its new
equilibrium temperature? Is this longer or shorter than the time it
would take Mars or Mercury to reach their respective equilibrium
temperatures?
b) What radiative forcing does this change correspond to?
c) Approximately how much warming would this brightening
eventually cause?
d) How would the calculated radiative forcing change if the
brightening takes place over 1,000 years instead of
instantaneously?
7. Explain why water-vapor changes are considered a feedback and
not a forcing.
8. The albedo changes from 0.3 to 0.31 on the Earth. What is the
radiative forcing associated with this change?
9. If doubled carbon dioxide has a radiative forcing of 4 W/m2, how
much of a change in albedo is required to completely cancel a
doubling of carbon dioxide on the Earth (put another way, how much
of a change in albedo is required to generate a radiative forcing of
−4W/m2)?
237

10. Imagine that, in addition to the fast feedbacks discussed herein,
there was also a fast negative “flower” feedback like that described in
this chapter (as the planet warms, white flowers prosper while black
ones die out), and that it had a magnitude g = −0.3. Estimate the
Earth’s climate sensitivity.
11. Assume that the Earth has warmed by 5°C since the last ice age,
and the change in radiative forcing over that time was +6.7 W/m2. On
this basis, calculate the climate sensitivity.
a) Express the climate sensitivity in °C/(W/m2).
b) Express the climate sensitivity in degrees Celsius per doubled
CO2.
12. In the northern hemisphere, Ein maximizes on June 21, when the
sun is most directly overhead. You might therefore expect
temperatures to be highest on that day. But for the U.S. Gulf Coast,
temperatures do not reach their hottest temperatures until mid-August
– several months after the maximum in Ein. Why?
1 I am implicitly assuming here that energy in always equals energy out
for the atmosphere. In more technical terms, this means that the heat
capacity of the atmosphere is zero. This is not a bad assumption
because, while not zero, the heat capacity of the atmosphere is indeed
much smaller than the heat capacity of the rest of the climate system.
2 The exact amount of warming depends on what we assume for
emissions of non-greenhouse-gas forcers, such as that from aerosols.
3 For somewhat technical reasons, radiative forcing is usually defined as
the change in energy balance at the tropopause. We ignore this technical
point here.
238

7
Why is the climate changing?
◈
In Chapter 2, we detailed the overwhelming evidence that the Earth’s
climate is changing – evidence so overwhelming, in fact, that few dispute
this anymore. Instead, the most heated argument is over the cause of the
warming: Is it caused by human activity, or is it natural? In this chapter,
we address this question.
Our strategy here is to examine the mechanisms that have changed
climate in the past and test each of them to determine if they could be the
cause of the recent warming. You will see that a careful assessment of all
of the possible causes yields the conclusion that the most likely
explanation for the recent warming is the increase in greenhouse gases in
our atmosphere, which we learned in Chapter 5 is due to human activity.
239

7.1 The first suspect: Continental drift
As you probably know, the Earth’s continents are moving. Not fast, mind
you – they move at about the rate that your fingernails grow – but over
tens of millions of years, this continental drift can substantially alter the
arrangement of the continents across the Earth’s surface. Such changes can
lead to large changes in the climate through several mechanisms.
For example, the location of continents determines whether ice sheets
form. The most important requirement for growth of an ice sheet is
summer temperatures cool enough that snow falling during the winter does
not melt during the following summer. This is most favorable for land at
high latitudes, which get the least sunlight.
Ice sheets matter to the climate because they reflect sunlight, so the
formation of an ice sheet increases planetary albedo, thereby increasing the
reflection of solar radiation back to space and cooling the planet. And the
loss of an ice sheet will warm the climate through the same mechanism. In
addition, the location of the continents determines the ocean circulation.
The oceans carry huge amounts of heat from the tropics to the polar
regions, so changing the ocean circulation can therefore alter the relative
temperatures of the tropics and polar regions. A good example of this
happened 30 million years ago when the Antarctic Peninsula separated
from the southern tip of South America, opening the Drake Passage. This
isolated Antarctica and allowed winds and water to flow unhindered
around it. This intense circumpolar flow reduced the transport of warm
water and air from the tropics to the South Polar region, cooling the
Antarctic and helping build the Antarctic ice sheet.
240

Continental drift can also indirectly affect the climate by regulating
atmospheric carbon dioxide. As I discussed in Chapter 5, carbon dioxide is
slowly removed from the atmosphere by chemical weathering, which
occurs when atmospheric carbon dioxide dissolves in rainwater and then
reacts with sedimentary rocks. The movement of the continents can change
the pattern of rainfall and expose new rock to the atmosphere, changing
the locations and rate of chemical weathering – thereby altering the
amount of carbon dioxide in the atmosphere. For example, 40 million
years ago the Indian subcontinent collided with Asia, forming the
Himalayas and the adjacent Tibetan Plateau. This change in surface
topology led to changing wind patterns, bringing heavy rainfall onto a vast
expanse of newly exposed rock in these features. The resultant chemical
weathering drew down atmospheric carbon dioxide over a period of tens of
millions of years. This is one of the reasons the planet has been generally
cooling over the past 40 million years (as shown in Figure 2.11)
Thus, continental drift can indeed affect the climate. But could it be
responsible for the rapid warming of the past few decades? The answer is
no. Because the movement is so slow, it takes millions of years for
continental movement to cause significant climate change. Continental
movements simply cannot significantly modify the climate over decades or
centuries.
241

7.2 The Sun
As we explored in Chapter 4, one of the factors that controls our climate is
the solar constant S. So if the Sun brightens or dims, we expect the climate
to respond by warming or cooling. It is therefore reasonable to ask if the
recent warming of the climate can be explained by an increase in the
brightness of the Sun.
It is well known that the Sun’s output varies on many time scales. For
example, solar physicists believe that, over the Sun’s 5-billion-year life, as
the Sun burned hydrogen and produced helium, the rate of fusion in the
Sun has increased as the buildup of helium increased in the density of the
Sun’s core. This has caused the Sun to become about 30 percent brighter
over this time.
Since the late 1970s, instruments on satellites have been accurately
measuring the solar constant; the measurements are plotted in Figure 7.1.
Over this period, there is a clear eleven-year cycle, during which the solar
constant varies by about 0.1 percent (about 1.3 W/m2). This cycle has little
influence on our climate, though. To understand why, imagine putting a
pot of water on the stove and then turning the burner beneath it on and off
each second. If you measured the temperature of the water, you would not
see it changing much each second. The thermal inertia of the water keeps
the temperature from varying quickly, so it does not respond to rapid
changes in heating. In our climate system, the thermal inertial of the ocean
prevents the climate from significantly responding to eleven-year solar
variations, just as it prevents the climate from instantly warming up in
response to increased radiative forcing (this was discussed in Section 6.1).
242

Figure 7.1 Percentage change in monthly values of energy in (Ein) due
to changes in the brightness of the Sun, based on the analysis of Frölich
and Lean. Seasonal changes in the Earth-Sun distance have been
removed (adapted from Forster et al., 2007, figure 2.16).
In order for the Sun to be responsible for the recent warming, there
would need to be a sustained, long-term increase in the solar constant over
the past few decades. The measurements show no evidence of this.
Another reason to discount the Sun as an explanation is that an increase in
solar output would warm the entire atmosphere. This is not happening.
Rather, measurements from weather balloons and satellites show that the
stratosphere (the region of the atmosphere beginning at an altitude of 10
km or so) has cooled over the past few decades. Thus, we can conclude
with high confidence that the rapid warming of the past few decades is not
caused by a brightening of the Sun.
The Sun’s influence on climate prior to the middle of the twentieth
century is more difficult to determine because there were no satellite
measurements of the solar constant. Instead, the Sun’s output for this
period must be inferred indirectly from other measurements, such as the
number of sunspots, which people have counted for many hundreds of
years, or from chemical proxies such as the carbon-14 content of plant
material. The most recent analyses of these records suggest that the Sun
has brightened over the past few hundred years, and this can potentially
243

explain some of the gradual warming of the eighteenth, nineteenth, and
early twentieth centuries. As I discussed in Chapter 6, this has led to a
positive radiative forcing with an estimated magnitude of +0.05 W/m2,
which is small compared to net radiative forcing from human activities
(+2.3 W/m2).
244

7.3 The Earth’s orbit
The solar constant is determined not just by the energy emitted by the Sun
but also by the Earth-Sun distance. If, for example, the Earth moved closer
to the Sun, then the solar constant would increase even if the brightness of
the Sun did not change.
This is relevant because the Earth’s orbit is not a perfect circle: It is an
ellipse whose eccentricity – the ratio of the length of the ellipse to the
width – varies with time. Over the course of 100,000 years or so, the orbit
cycles between an orbit that is slightly more eccentric (more elliptical) and
one that is less eccentric (more circular) (Figure 7.2).1 As the orbit
becomes more elliptical, the average Earth-Sun distance increases and the
average amount of solar energy falling on the Earth decreases. For the
Earth’s orbit, this causes the annual average solar constant to vary by
approximately 0.5 W/m2 over the 100,000-year cycle.
Figure 7.2 Schematic illustrating how the eccentricity of the Earth’s
orbit (how elliptical it is) varies with a period of 100,000 years or so.
Other aspects of the Earth’s orbit can also vary, such as the timing of
the closest approach of the Earth to the Sun (also known as the perihelion).
Today, the Earth is closest to the Sun during January, when it is wintertime
in the northern hemisphere. Over the next 23,000 years, the date of closest
245

approach will cycle through the entire year. In roughly 11,500 years, the
Earth will be closest during July, and in 23,000 years it will again be
January.
Another important variation is the tilt of the Earth (also known as the
obliquity). Today, the Earth’s spin axis is tilted 23.5° from vertical (Figure
7.3). However, over the next 41,000 years, the Earth’s tilt will complete a
cycle through a range of tilt angles from 22.3° to 24.5°.
Figure 7.3 Schematic illustrating how the obliquity of the Earth (the tilt
of the spin axis away from a line perpendicular to the orbital plane, the
plane defined by the Earth’s orbit) varies with a period of 41,000 years.
Changing the date of closest approach to the Sun or the tilt of the
Earth does not change the Earth-Sun distance, so it does not change the
solar constant. Rather, these changes change how sunlight is distributed
over the planet, in both latitude and season. For example, increasing the tilt
of the planet increases the amount of sunlight hitting the polar regions and
decreases the amount hitting the tropics. Such changes can alter the
climate.
246

We see in the paleoclimate record nearly perfect agreement between
the ice-age cycles (Figure 2.13) and the variations in the Earth’s orbit. As I
discussed earlier in this chapter, the growth of big, continental-scale ice
sheets, such as existed during the last ice age, is determined by high-
latitude summertime temperatures – because this determines whether snow
that falls during the winter survives the subsequent summer. Orbital
variations regulate how sunlight is distributed over the planet and over the
seasons, so they play a key role in regulating these temperatures. These
orbital variations and the climate effects that follow are often referred to as
Milankovitch cycles, after Serbian mathematician Milutin Milankovitch,
who was the first one to recognize that the ice-age cycles corresponded to
variations in the Earth’s orbit.
But while these orbital variations are critical in ice ages, are they
responsible for the warming of the past few decades? They are almost
certainly not. These orbital variations are so slow that it takes thousands or
tens of thousands of years to make any significant change in the amount of
or distribution of incoming sunlight. The warming of the past century has
been much too fast to be caused by these slow orbital variations. The
warming must be due to other causes.
247

7.4 Internal variability
Changes in the output of the Sun or in the Earth’s orbit are examples of
forced variability: changes in the Earth’s climate in response to a radiative
forcing. However, the Earth’s climate system is so complex that it can also
change without any external factors driving it. Such changes, which are
caused by the internal physics of the system rather than external changes in
the planet’s energy balance, are often referred to as internal variability.
A good example of internal variability that you can put on your desk
is the “drinking bird,” a toy in which a toy bird oscillates between standing
straight up and rotating over and sticking its beak into a glass of water.2
The drinking bird relies on internal physics to drive this oscillation; its
periodic behavior is not externally forced.
The best-known example of internal variability in our climate is the
El Niño/Southern Oscillation (ENSO). El Niño events, which make up the
warm phase of ENSO, occur every few years and last a year or so. During
these events, the Earth warms several tenths of a degree Celsius. El Niño’s
opposite is La Niña, and during those events the Earth cools several tenths
of a degree. These ENSO events cause a temporary temperature change
every few years but no long-term changes in the climate. Figure 2.4, for
example, shows the dramatic warming that the Earth experienced during
the El Niño of 1998 as well as the cooling experienced during the La Niña
of 2008. In fact, many of the short-term variations in the temperature
record can be traced back to ENSO events.
ENSO is the dominant and best-known source of internal variability
in the climate system. However, other modes of variability, with names
like the Pacific Decadal Oscillation (PDO) or the Atlantic Multi-decadal
248

Oscillation (AMO), are thought to exist, although they are less well
understood. So the relevant question is this: Could the warming of the past
few decades be due to one of these modes of internal variability? We can
rule out ENSO as the cause because that cycle lasts a few years at most, so
it could not cause a warming trend lasting several decades.
The proxy data on climate variation before the past two centuries can
help answer whether another mode of internal variability – one that lasts
for decades, not years – may be responsible. Human activities probably
had minimal impact on climate before 1800, so climate paleoproxy data
before that time should provide a good picture of recent natural climate
variability. As we could see in Figure 2.15, the record between 1000 AD
and 1800 AD shows nothing similar to the rate and magnitude of warming
of the twentieth century. Thus, the paleoproxy data do not support internal
variability as a cause of the recent warming.
We can also gain insight into natural climate variability by running
climate models without any human greenhouse-gas emissions. In these
simulations, climate models exhibit variations in global average
temperature from year to year and decade to decade that are similar to
those seen in the climate proxy data before about 1800, but they produce
nothing resembling the rapid warming of the past century. The final
argument against internal variability is that no one has identified any
physical mechanism that would explain the warming.
Ultimately, we cannot exclude internal variability the same way we
can definitively exclude, say, a brightening Sun. However, there is
basically no evidence supporting this explanation, either. So internal
variability is like a suspect in a criminal investigation who has no alibi but
for whom there is also no evidence linking him to the crime. You cannot
exclude the suspect, but you would be hard-pressed to convict him based
only on a lack of an alibi.
249

7.5 Greenhouse gases
The last potential explanation for the recent warming is the increase in
greenhouse gases in our atmosphere. Chapter 4 gave the physical
explanation for why an increase in greenhouse gases would be expected to
warm the planet, and Chapter 5 discussed in detail why this increase is
almost entirely due to human activities.
This physics is neither new nor complex; it was first recognized in the
early nineteenth century, and our knowledge of it has been refined by
nearly two centuries of work by thousands of scientists, including
luminaries such as Fourier, Tyndall, and Arrhenius. In agreement with
these simple physical arguments, the geologic record over the past 500
million years shows a strong correlation between temperature and
atmospheric carbon dioxide. An example of this can be found in Figure
7.4, which shows periods of widespread glaciation (indicated by gray bars
in the figure) when carbon dioxide was low and less ice when carbon
dioxide was high.3 Since ice extent is a proxy for temperature, we can
conclude that variations in carbon dioxide and temperature have generally
been associated with each other for much of the Earth’s history.
250

Figure 7.4 Atmospheric carbon dioxide over the past few hundred
million years. The gray bars indicate times when ice existed on the
planet (based on Royer, 2006, figures 1 and 2).
An aside: How does science deal with outliers?
Figure 7.4 also provides a good example of why climate science is
hard. Although most glaciations are associated with low carbon
dioxide, the eagle-eyed among you will notice that approximately
450 million years ago there was a glaciation when carbon dioxide
levels were greater than 5,000 ppm. Such a point is known as an
outlier – a point that does not agree with the rest of the data. So-
called climate skeptics might take this single point and argue that it
disproves the connection between climate change and greenhouse
gases. Is that a reasonable conclusion? What would a scientist
think about this outlier?
There are several possible explanations for the outlier. First,
the theory connecting carbon dioxide with climate may indeed be
wrong, as the skeptic suggests. Second, the data may be wrong –
perhaps there was no glaciation, or maybe carbon dioxide was
really much lower than suggested by the proxy data. After all, we
are trying to infer the conditions on the planet nearly half a billion
years ago, and there are lots of ways that the proxy data could
mislead us. Finally, both the data and greenhouse gas theory could
be right: there may have been something else offsetting the
warming from carbon dioxide. For example, massive volcanism
could have injected enough aerosols into the atmosphere to lead to
low temperatures despite high abundances of carbon dioxide.
251

In his seminal work, The Structure of Scientific Revolutions,
Thomas Kuhn described how incorrect scientific theories
accumulate anomalies – places where the observations do not
match theory. These anomalies accumulate until there are so many
that the theory is simply no longer tenable, and a scientific
revolution overthrows the old theory and replaces it with a new
one.
For example, at the beginning of the twentieth century,
physics was in trouble. The classical theory of physics could not
explain several well-validated observations, including the T4
dependence of blackbody radiation (discussed in Chapter 3),
atomic and molecular spectra, and the photoelectric effect.
Eventually, it became apparent that classical physics simply did
not work at the atomic level, and a scientific revolution occurred.
What emerged was a new paradigm, in which quantum mechanics
ruled small, atomic domains, and classical physics ruled our
macroscopic, everyday world.
It is important to recognize that outliers occur in all fields, not
just climate science. For example, you can find – contrary to
expectations – people who smoke four packs of cigarettes each day
yet who live to be ninety years old. Such anomalies frequently
allow scientists to refine and extend their theories: Given that
smoking causes cancer, why are some people less susceptible than
others? Is it just luck, or is there a physiological basis?
Importantly, though, the existence of some smoking anomalies
does not cause scientists to reject the underlying idea that smoking
is bad for your health.
The question that each scientist must ask individually, and the
scientific community must ask collectively, is whether a particular
252

theory has accumulated enough anomalies that it is no longer
tenable. At present, there are not enough anomalies like the
glaciation 450 million years ago to reject the dominant theory that
greenhouse gases play a major role in determining our climate. But
scientists are always looking for new anomalies – and if enough of
them accumulate, eventually this theory of climate will be replaced
by another one.
From a practical standpoint, though, no one expects that to
occur. The theory that carbon dioxide exerts a strong influence on
climate is so successful in predicting so many aspects of our
climate that it is quite unlikely that the theory will turn out to be
substantially wrong. This is akin to our views on smoking and lung
cancer. Although it is possible that future research may disprove
the link, it is a very, very unlikely eventuality.
Another example of the correlation between greenhouse gases and
temperature is an event roughly 55 million years ago known as the
Paleocene-Eocene Thermal Maximum or PETM, which was discussed in
Section 2.2.2. This event began with a massive release of either carbon
dioxide or methane, which in turn led to an increase in the Earth’s global
average temperature of 5–9°C over the following few thousand years (top
panel of Figure 7.5). The mass of carbon was so immense that when it
dissolved into the oceans, the oceans became significantly more acidic (as
I discussed in Chapter 5, carbon dioxide forms carbonic acid after it
dissolves in water). This in turn dissolved calcium carbonate (the material
that makes up shells) in the sediments at the bottom of the ocean (bottom
panel of Figure 7.5).
253

Figure 7.5 Temperature during the Paleocene-Eocene Thermal
Maximum as a function of time from various sites (top panel); calcium
carbonate content of ocean sediments (bottom panel)
(adapted from Jansen et al., 2007, figure 6.2).
The temperatures remained elevated for 100,000 years or so, which is
about the length of time it takes the carbon cycle to fully remove the
carbon from the atmosphere. Interestingly, the mass of carbon released, a
few thousand gigatons, is comparable with the amount contained in all of
the Earth’s fossil fuels. Thus, the PETM is frequently viewed as a good
analog to what will happen if humans burn all of the fossil fuels over the
next few centuries. An important difference, however, is that humans are
on a pace to release the carbon over several hundred years, whereas it was
released during the PETM over several thousand years. Thus, we can
expect even more rapid warming over the next few centuries than that
experienced during the PETM, which was a period of – geologically
speaking – very rapid warming.
254

The association between carbon dioxide and temperature is even
clearer over the past few hundred thousand years. Figure 2.11 shows how
carbon dioxide and temperature varied in lock step as the Earth cycled
between ice ages and warm interglacials. The association between
temperature and carbon dioxide, however, is a bit more complicated than
the plot may at first suggest. There is strong evidence that ice-age cycles
are initiated by small variations in the Earth’s orbit (as discussed in Section
7.3). However, the changes in sunlight falling on the Earth in response to
these slight orbital changes are too small to explain the wide temperature
swings during ice-age cycles. Something must be helping the orbital
variations produce the observed variations.
What is missing is carbon dioxide. The small initial warming from
orbital variations leads to increased levels of carbon dioxide through
mechanisms that have not yet been unambiguously identified. The increase
in carbon dioxide leads to further warming. In other words, the orbital
variations are the forcing, and carbon dioxide is acting here as a feedback
that amplifies the small initial warming from the forcing.
We are not exactly sure what process releases carbon to the
atmosphere as the climate warms during ice-age cycles. As we learned in
Chapter 5, the two biggest (non-human) sources of carbon for the
atmosphere are the land biosphere and ocean. For both of these reservoirs,
there are plausible mechanisms that could explain why warmer
temperatures would release carbon to the atmosphere, but our confidence
in the details of the carbon cycle’s response to climate change is low.
An aside: A skeptical argument
During the ice ages, carbon dioxide began rising after the
temperature. This proves that carbon dioxide responds to
255

temperature, and not the other way around. Ergo, carbon dioxide
cannot be causing the present-day warming.
The problem with this argument is that it misunderstands the
difference between a climate forcing and a feedback. When you
have a forcing, such as the Sun getting brighter or the addition of
greenhouse gases to the atmosphere, the temperature change
occurs after the forcing is applied to the climate system.
A feedback, however, is more complicated. There is an initial
temperature change, followed by the feedback mechanism,
followed by additional warming. For the ice-albedo feedback
diagramed in Figure 6.8, there is an initial warming, followed by a
loss of ice, followed by increased absorption of solar energy,
followed by more warming. In this case, warming is first, then the
ice melts. In this case, it is wrong to conclude that, because the
temperature change occurs first, the melting ice has no effect on
the climate. It does.
The key point here is that carbon dioxide is presently a
climate forcing – humans are adding it to the atmosphere, and that
is causing warming. During the ice-age cycles, however, carbon
dioxide acted as a feedback – a warming planet led to the release
of carbon dioxide, which then caused additional warming.
A final link between greenhouse gases and climate comes from
climate models. Simulations of the twentieth century by climate models
that exclude the observed increase in greenhouse gases fail to simulate the
increase in temperature over the second half of the twentieth century. This
can be seen in Figure 7.6: the model run in Figure 7.6a includes natural
forcings – primarily changes in the solar constant and volcanoes – but no
256

human impact on climate. This calculation reproduces many of the bumps
and wiggles in the record, showing that these are not due to human
activity. But this simulation completely fails to capture the rapid warming
that began around 1960.
Figure 7.6 Global mean surface temperature anomalies from the surface
thermometer record (lighter curves), compared with a coupled ocean-
atmosphere climate model (darker curves). The model includes (a) only
nonhuman natural climate forcing, in particular solar and volcanic
effects, and (b) natural forcing and human greenhouse-gas emissions,
aerosols, and ozone depletion. Anomalies are measured relative to the
1901–1950 mean
(the source is figure 3.12 of Dessler and Parson, 2010, which is an
adaptation of figure TS.23 of Solomon et al., 2007).
The model run in Figure 7.6b includes both natural effects as well as
the effects of human activities – mainly greenhouse-gas emissions but also
257

increases in aerosols and decreases in stratospheric ozone. This model
captures the rapid warming since 1960 that the model with only natural
forcing fails to simulate. This suggests that human greenhouse-gas
emissions, volcanic effects, and solar effects have all contributed to global
temperature changes of the past century but that greenhouse-gas emissions
are responsible for most of the rapid late-twentieth-century warming.
258

7.6 Putting it all together
As we learned in Chapter 2, the Earth’s climate has varied more or less
continuously for the past several hundred million years, and probably for
the entire history of the planet. Obviously, most of these variations have
nothing to do with human activities.
Thus, when we consider the recent warming, the first thing we must
do is investigate whether today’s warming is due to natural variations. In
so doing, most natural explanations can be decisively eliminated (e.g.,
continental drift, orbital variations, variations in the output of the Sun).
Internal variability cannot be eliminated, but there is little evidence to
support that as an explanation.
In contrast, there is overwhelming evidence that the increase in
greenhouse gases is the cause of the recent warming. There is strong
theoretical evidence that greenhouse gases warm the planet, including the
simple arguments detailed in Chapter 4 and the more sophisticated
calculations of climate models. There is also observational evidence that
carbon dioxide has played a key role in our climate over the past 500
million years.
Taken together with the lack of a competing hypothesis, the totality of
evidence that carbon dioxide is the main cause of the recent warming
makes a compelling case. Reflecting this, the 2013 report of the
Intergovernmental Panel on Climate Change came to the following
conclusion:
It is extremely likely that more than half of the observed increase in
global average surface temperature from 1951 to 2010 was caused by
259

the anthropogenic increase in greenhouse gas concentrations and
other anthropogenic forcings together.
Note that this statement is carefully caveated in three ways. The first is the
phrase more than half. This makes it clear that greenhouse gases are not
the only factor that can influence the climate. As we have explored in this
chapter, there are other factors that can influence the climate, and some of
these (e.g., solar variations, internal variability) may have been minor
contributors to the recent warming. However, increases in greenhouse
gases are responsible for the majority of the observed warming.
The second caveat is the specified time period, 1951–2010. It is only
during this period that our observations are sufficient to rule out alternative
explanations for the warming. It is certainly possible that greenhouse gases
were the dominant cause of the warming prior to this period, but that
cannot be proven to the high standards required by the scientific
community. This leaves the pre-1951 warming to be, officially at least,
unattributed.
The third caveat concerns the words “extremely likely.” The
Intergovernmental Panel on Climate Change uses a set of carefully defined
terms to express confidence. In the parlance of the IPCC, extremely likely
denotes a confidence of 95 percent. This acknowledges that the
mainstream scientific view may indeed be wrong (unlike the evidence that
the globe is warming, which the IPCC describes as unequivocal).
However, the chance is small – about one in twenty.
An aside: Evolution of the IPCC’s statements on warming
The IPCC has produced five major assessments of the science of
climate change since 1990 – one about every six years. Below you
260

can see the evolution of the IPCC’s key statement on the attribution
of the observed warming to humans – from a relatively weak
statement in 1990 to a strong attribution in 2013.
1990: The size of the observed warming “is broadly
consistent with predictions of climate models, but it is also of
the same magnitude as natural climate variability. Thus the
observed increase could be largely due to this natural
variability.” This statement reflects the fact that climate
science was in its infancy at that time. Satellite measurements
had been available for little more than a decade, computers
were slow, and there were few climate scientists in the world
working on the problem. As a result, it was not possible at
that time to demonstrate a clear human impact on the climate.
1995: “The balance of evidence suggests a discernible human
influence on global climate.”4 In the time since the 1990
report, many advances had occurred in climate science. Most
of the new evidence suggested that humans were having some
effect on the climate, although an argument could still be
made that the warming was mostly natural. The statement did
not include any quantification of the human influence.
2001: “Most of the observed warming over the last 50 years is
likely to have been due to the increase in greenhouse gas
concentrations.” This statement makes a much more definitive
statement about the role of humans than the earlier statements
did. This reflected improvements in our observations of the
climate system, improvements in computers and climate
models, and advances in our theoretical understanding of the
261

planet. The word likely denotes a 66 percent chance (two out
of three) that the statement is true.
2007: “Most of the observed increase in global average
temperatures since the mid-twentieth century is very likely
due to the observed increase in anthropogenic greenhouse gas
concentrations.” The 2007 statement is essentially identical to
the 2001 statement, except it projects a higher level of
confidence by using the words very likely, which denotes a
nine out of ten chance that the statement is true.
2013: “It is extremely likely that more than half of the
observed increase in global average surface temperature from
1951 to 2010 was caused by the anthropogenic increase in
greenhouse gas concentrations and other anthropogenic
forcings together.” The 2013 statement continues the growth
in our confidence in the statement: likely in 2001 to very likely
in 2007 to extremely likely in 2013.
What is most striking, to me at least, is how climate science has
actually not changed much over the years and decades. What
Arrhenius thought in 1896, what scientists studying the climate
thought in the 1950s, 1960s, and 1970s, and what the IPCC report
described in 1990 is basically what we think in 2013. The only
difference is that our confidence in this understanding has vastly
improved.
The overall stability of climate science should provide us with
great confidence in it. This is because important scientific ideas are
constantly retested by scientists, so the longer an idea survives, the
more likely it is to be correct. And as a perusal of the IPCC reports
262

shows, most of the major claims of climate science have indeed
survived a long time.
Any analysis of the cause of the recent warming should also be clear
about what is not evidence that greenhouse gases are the primary cause of
the recent warming. The case for greenhouse gases is not built on the
argument that the present temperature of the Earth is exceptional. In fact,
we know that the Earth has been much warmer than it is today. As Figure
7.4 shows, over much of the past 500 million years the Earth was so warm
that there was no ice anywhere on the planet. Nor is the case for
greenhouse gases built on the argument that the rate of today’s warming is
exceptional. It may be that today’s warming is indeed without precedent,
but we simply do not have the data covering the entire Earth’s history to
prove that. Nor is the case built on the occurrence of extreme events, such
as the extreme Atlantic hurricane season of 2005, Superstorm Sandy in
2012, or the 2003 European heat wave. Rather, the case for greenhouse
gases is built on a thorough examination of all of the possible
explanations. The fact that the increase in greenhouse gases can explain
the warming, combined with the distinct lack of any legitimate competing
theory, leads scientists to conclude that it is extremely likely that that
greenhouse gases are indeed responsible.
263

7.7 Chapter summary
To determine a cause for the present-day warming, we examine all
of the natural processes that are capable of changing our climate.
Among these are continental drift, variations in the Sun, and orbital
variations, which can all be decisively rejected as explanations for
the present-day warming. Internal variability, such as El Niño
cycles, cannot be definitively eliminated as a significant cause of
long-term warming, but there is also no evidence that it is.
There is abundant evidence that the increase in greenhouse gases,
which is due primarily to human activities, can explain the present-
day warming. There is strong theoretical evidence that greenhouse
gases warm the planet, including the simple arguments detailed in
Chapter 4. And sophisticated calculations by climate models are
only capable of reproducing the warming of the past half-century if
the increase in greenhouse gases is included. There is also strong
observational evidence that carbon dioxide has played a key role in
our climate over the past 500 million years.
On the basis of this evidence, the IPCC concluded in its 2013
report that “It is extremely likely that more than half of the
observed increase in global average surface temperature from 1951
to 2010 was caused by the anthropogenic increase in greenhouse
gas concentrations and other anthropogenic forcings together.”
264

Terms
Continental drift
Eccentricity
Forced variability
Internal variability
Milankovitch cycles
Obliquity
Outlier
Perihelion
265

Additional reading
The Working Group I report of the IPCC’s Fifth Assessment describes, at
varying levels of detail, the evidence attributing the recent warming to
humans. For the most detail, see Chapter 10 of the report. For a less
detailed overview, see Section TS.3 of the Technical Summary. And, for a
short, high-level discussion, see Section D.3 of the IPCC’s Summary for
Policymakers (you can download all of these at
www.ipcc.ch/report/ar5/wg1/).
SkepticalScience.com has several useful write-ups that summarize the
evidence that humans are responsible for most of the recent warming
(www.skepticalscience.com/its-not-us-basic.htm,
www.skepticalscience.com/empirical-evidence-for-global-warming.htm).
See www.andrewdessler.com/chapter7 for additional resources for
this chapter.
266

http://www.ipcc.ch/report/ar5/wg1/

http://www.skepticalscience.com/its-not-us-basic.htm

http://www.skepticalscience.com/empirical-evidence-for-global-warming.htm

http://www.andrewdessler.com/chapter7

Problems
1.
a) List all of the physical processes that can alter the climate.
b) For all processes in part (a) except greenhouse gases, explain
why they are unlikely to be the cause of the warming over the past
few decades.
c) List the evidence that greenhouse gases are responsible for the
recent warming.
2. What did the IPCC say in its 2013 report about whether humans are
causing climate change? What are the three caveats in the statement?
3. Why are feedbacks (e.g., increases in water vapor) not discussed as
potential causes of climate change?
4. Explain the physical mechanism for the occurrence of ice ages.
Make sure you explain the role of carbon dioxide and its timing with
respect to the temperature change.
5. Critique this statement: “It is clear that it was warmer around 1000
AD, during the Medieval Warm Period, than it is today. Therefore,
humans cannot be causing today’s warming.” Assume that the claim
that the Medieval Warm Period was warmer than today is correct (it
may have been, but it is debatable). Is this argument correct? Why or
why not?
6. What are the three ways that the Earth’s orbit varies? How does
each variation affect the climate?
267

7. Explain how the Paleocene-Eocene Thermal Maximum provides
support for the claim that today’s warming is caused by humans.
8. How does continental drift affect our climate?
1 When we specify the length of a cycle (e.g., the 100,000-year
eccentricity cycle), this is the length for the eccentricity to execute one
complete cycle. This means that it takes 50,000 years for the
eccentricity to vary from its maximum value to its minimum, and
another 50,000 years to return to its starting value.
2 See www.andrewdessler.com/chapter7 for links to videos of the
drinking bird.
3 Over this same time, the Sun brightened by 6 percent or so. Thus, the
climate associated with a certain level of carbon dioxide a few hundred
million years ago would be cooler than the climate would be for the
same amount of carbon dioxide today. This explains why glaciations
were occurring hundreds of millions of years ago with carbon dioxide
abundances higher than today’s.
4 The evolution of this statement is documented in a great article by
Houghton (2008).
268

http://www.andrewdessler.com/chapter7

8
Predictions of future climate
change
◈
In Chapter 6, we discussed the concept of radiative forcing, which is an
imposed change in planetary energy balance. In response, the planet’s
temperature adjusts so as to restore energy balance. Thus, if we can predict
how radiative forcing will evolve in the future, we can then estimate how
much climate change we will experience.
Predicting future radiative forcing basically comes down to predicting
how much greenhouse gas and aerosols will be emitted into the
atmosphere each year from human activities. Such projections, known as
emissions scenarios, therefore form the backbone of our predictions of
climate change. In this chapter, I describe how they are constructed and
what they tell us about our future climate.
269

8.1 The factors that control emissions
At its simplest, the amount of greenhouse gas released by a society is
determined by the total amount of goods and services consumed by that
society. This is true because the production of any good or service – be it a
car, an iPhone, a university lecture, a cheeseburger, or an hour of tax
consulting – requires energy. And energy is mostly derived from the
combustion of fossil fuels, which leads to the release of carbon dioxide to
our atmosphere. The emissions of other greenhouse gases and aerosols
also generally scale with the amount of consumption, although the causal
linkages may not be as direct.
The total value of goods and services produced by an economy is
known as the gross domestic product, abbreviated GDP. Thus, total
emissions by a society are basically set by that society’s GDP. If the GDP
doubles, then we expect emissions to double, as long as everything else
remains the same. This strong link between GDP and emissions can be
seen during recessions. For example, during the severe economic
downturn of the late 2000s, global carbon emissions posted their biggest
drop in more than forty years as the global recession froze economic
activity and slashed energy use around the world.
Rather than consider GDP as a whole, it is useful to break it into the
product of two factors: population and affluence. It should be obvious that
GDP scales with population. Every person in a society consumes goods
and services, so if the population doubles (and everything else remains the
same), then total GDP will also double. Emissions will therefore also scale
with population – so emissions double if the population doubles.
270

In addition to the number of people, how rich each person is also
matters because, as people get richer, they consume more. To illustrate the
affect of affluence on GDP and emissions, consider the following three
families. The first is a family of four who live as subsistence farmers in
sub-Saharan Africa. This family lives in a small one-room house without
electricity or running water. They do not own a car and are too poor to buy
anything but the bare necessities of life. They farm by hand or with a draft
animal. Because the members of this family are so poor and consume so
little, they are responsible for little greenhouse-gas emissions.
Now consider a family of four near the bottom of the economic
spectrum in the United States. They live in an apartment and they own one
car. Their apartment is not air-conditioned; they own a television and one
or two heavy-duty electrical appliances, such as an oven. Compared with
the subsistence farming family in Africa, this family is far richer and
consumes far more and is therefore responsible for more greenhouse-gas
emissions.
Finally, consider an upper-class family of four in the United States.
This family lives in a 4,000-ft2 single-family house and owns three cars
(for the husband, wife, and a teenage child). The house has televisions in
almost every room, several computers, VCRs, game consoles, and a rich
assortment of electrical appliances. The family flies to several vacation
locations every year. Because of the significant consumption allowed by
their affluence, this family is responsible for more emissions than the
poorer U.S. family and many, many times the emissions of the subsistence
farming family.
This wealth effect leads to enormous disparities in emissions per
person. In the United States, emissions are about 5 tons of carbon per
person. Emissions in China are 1.7 tons per person – about one-third of
U.S. per capita emissions. However, China’s population is so large that
271

(8.1)
(8.2)
they nevertheless lead the world in total carbon emissions. Emissions in
Nigeria are 0.1 tons per person – about one-fiftieth of the United States –
reflecting the country’s poverty.
We need a third factor to convert a level of total consumption,
expressed in dollars, to greenhouse-gas emissions. This last factor relates
how much greenhouse gas is emitted for every dollar of consumption; it is
known as the greenhouse-gas intensity. Putting these all together, we can
now relate emissions to the factors that control it in a simple equation:
Here I represents the total emissions of greenhouse gases into the
atmosphere (these emissions then cause climate impacts, which is why
emissions are represented by the letter I); P is the population, A stands for
affluence, and T stands for greenhouse-gas intensity. Affluence A is GDP
per person – the average amount of goods and services each person
consumes – so the product of P and A is the GDP. The decomposition of
emissions into these factors is often referred to as the IPAT relation or the
Kaya Identity.
The greenhouse-gas-intensity term T can be usefully broken down as
the product of two terms:
EI stands for energy intensity – the number of joules of energy it takes to
generate one dollar of goods and services. The EI of an economy is
primarily determined by two factors. First is the mix of economic activities
that make up the economy. For example, it takes much more energy for a
steel mill to produce one dollar’s worth of steel than for a university to
produce one dollar’s worth of teaching. The steel mill must run blast
furnaces and other heavy equipment, whereas the university only requires
272

lighting, air-conditioning, computers, and the like. More generally,
industrial manufacturing has a higher energy intensity than white-collar
service-oriented activities. The more industrial manufacturing an economy
has, the higher its energy intensity.
The second factor in determining the energy intensity of an economy
is the efficiency with which the economy uses energy. For any economic
activity, there are usually several technologies to accomplish it. For
lighting, for example, there is the standard incandescent light bulb (the
kind with the filament) or the compact fluorescent light bulb. As described
in Chapter 3, incandescent light bulbs are dreadfully inefficient, requiring
60 W of power to produce the same light as a compact fluorescent light
bulb drawing 14 W. Both light-bulb technologies can light a room, but
they consume vastly different amounts of energy doing it. The trade-off is
that better technology is often more expensive. As a result, it takes a
certain level of wealth in order to adopt the most energy-efficient
technology, and the efficiency with which different countries utilize
energy can vary greatly.
CI in Equation 8.2 stands for carbon intensity – the amount of
greenhouse gas emitted per joule of energy generated – which reflects the
mix of technologies used to generate energy. Put another way, it is
determined by whether the economy uses coal, oil, gas, nuclear, wind,
solar, etc. to generate energy. Among fossil fuels, combustion of natural
gas (methane or CH4) produces the least carbon dioxide per joule of
energy generated. Thus, it has the lowest carbon intensity, which is one of
the reasons it is often considered to be the “greenest” of the fossil fuels.
Oil produces more carbon dioxide per joule than methane, so it has a
higher carbon intensity. The most carbon-intensive fossil fuel is coal – it
produces roughly twice the carbon dioxide per joule as methane – which
explains why many people who are concerned with our climate are
273

opposed to the construction of new coal-fired power plants. Energy
sources also exist that produce no carbon dioxide, such as hydroelectric,
nuclear, wind, and solar energy sources.
For a country such as France, which generates most of its electricity
from nuclear energy, the carbon intensity will be smaller than for a country
such as China or the United States, which both rely heavily on coal for
electricity.
An aside: Check the units!
One of the most powerful ways to check your work is to make sure
that the units in a problem work out. We do this now to close the
loop on our understanding of the factors that regulate carbon
emissions.
Population is obviously the number of people. Affluence is
dollars of GDP per person. The product of population and
affluence is therefore GDP, which has units of dollars:
Energy intensity has units of joules per dollar, and carbon intensity
has units of carbon dioxide emitted per joule. Greenhouse-gas
intensity is the product of energy intensity and carbon intensity,
and therefore it has units of carbon dioxide emitted per dollar:
Finally, the product of population, affluence, and technology has
units of carbon dioxide emitted:
274

275

8.2 How these factors have changed in
the recent past and how will they
change in the future
In the last section, emissions of carbon dioxide were deconstructed into the
controlling terms: population, affluence, energy intensity, and carbon
intensity. Let us look at how these terms have changed over the past few
decades and how they might change in the future.
276

8.2.1 Population
Population has been rapidly increasing for the past few centuries. It took
all of human history up to 1804 for the global population to reach 1 billion.
The 2-billion-people mark was reached 123 years later, in 1927, and the 3-
billion-people marker was reached 33 years later, in 1960. Since then,
world population has been increasing by 1 billion people every twelve to
thirteen years, reaching 6 billion in 1999 and 7 billion in 2011. Figure 8.1
shows that the population has increased by 80 percent over the past few
decades. Today, world population is increasing by roughly 200,000 people
per day, a population growth rate of approximately 1 percent/year. Most of
this growth is occurring in the developing world, where fertility rates
remain high.
Figure 8.1 Population, affluence, carbon intensity, and energy intensity
for the entire world, relative to values in 1970
(adapted from IPCC, 2007b, figure 2).
In estimating future population change, some of the controlling
factors are well known. Affluence, for example, strongly determines how
many kids a woman has, with the poorest countries having the highest
fertility rates. In extremely poor societies, children can be put to work at a
young age and are therefore a source of income. This is generally not the
277

case in rich countries, where children are a net drain on family resources
for many years (trust me on that). In addition, high rates of childhood
death in poor countries mean that parents must have many children to
ensure that some of them survive into adulthood. Improvements in health
care that occur as a society gets richer, however, mean that rich parents
can reasonably expect their children to survive into adulthood. The amount
of education that women receive is also a factor, with fertility rates
declining as women become better educated and good-paying jobs become
available to them as an alternative to child rearing. Our understanding of
these factors gives us some ability to predict future population.
However, some events that affect population are impossible to
predict. It is impossible to predict, for example, societal changes, such as a
future Pope suddenly embracing birth control, causing a fertility decline
among the roughly one billion Catholics. Or the occurrence of chance
events, like a nuclear war or the emergence of new diseases like AIDS,
which kill millions of people.
The lowest population scenarios predict population will peak at
around 9 billion in the middle of the twenty-first century, decreasing
thereafter and reaching 7 billion by 2100. Higher population scenarios
predict population increasing, more or less continuously, reaching 16
billion by 2100. Our best estimate is that world population will stabilize
during the twenty-first century around 10 or 11 billion.
278

8.2.2 Affluence
Figure 8.1 shows that affluence, measured as GDP per person, increased
by 80 percent over the past few decades of the twentieth century. My
personal experience backs this up. When I was a college student in the
early 1980s, I did not own a cell phone or a laptop or tablet computer, my
car had handcranked windows, did not have air conditioning or antilock
brakes or an airbag, and only had an AM radio. I did not own a TV or
videogame console. I had, in other words, a far lower standard of living
than most of today’s college students.
We can also see discrete political events in the affluence data, such as
the 1989 collapse of the Soviet Union and the associated political upheaval
in Eastern Europe, as well as various recessions. Moreover, in the past
decade, the remarkable economic growth of China (with affluence growing
at 10 percent/year or so) and other large developing countries has played a
key role in driving global growth of consumption.
In the future, factors such as the level of education in the population,
rule of law, free trade, and access to technology will be key in determining
how fast affluence grows. In general, economic growth rates are highest
for countries making a transition out of poverty and into the group of rich
countries of the world. For example, economic growth was fastest in the
United States in the late nineteenth century, in Japan after World War II,
and in China today. Growth rates are lower for large, advanced societies.
Based on these factors, expert predictions are that affluence will increase
over the twenty-first century at 2 to 3 percent/year for developing
countries and 1 to 2 percent/year for industrialized countries.
279

8.2.3 Technology
The first part of the technology term, the energy intensity term, has
decreased over the past century as our society has developed more
efficient ways to use energy (Figure 8.1). Some of this increasing
efficiency has been driven by market forces: Because energy costs money,
a more energy-efficient piece of equipment or process will reduce costs,
which consumers want. As a result, just about everything you buy today is
more energy efficient than the comparable version of a few decades ago.
Much of this increase in efficiency is incremental, meaning that each new
generation of a particular piece of equipment uses slightly less energy than
the previous version. Sometimes, however, there is a revolution in
technology that greatly reduces energy consumption. A good example is
the revolution in lighting technology we are now experiencing. As the
world switches from incandescent bulbs to compact fluorescent bulbs and
LED bulbs, the amount of energy being consumed by lighting will
experience a substantial one-time drop.
Changes in the mix of goods and services produced by the world’s
economy has also led to decreases in energy intensity. Over the past
century, the fraction of the world economy based on energy-intensive
heavy industry and manufacturing has declined, while the fraction based
on services has increased.
Overall, energy intensity has at times decreased as fast as 2
percent/year, but the periods of fastest decreases occurred during periods
of rapid economic shifts or as responses to energy price shocks. More
typical values for the twentieth century were decreases of 1 percent/year. It
is likely that this rate of decrease can be sustained over the coming
century.
280

Figure 8.1 also shows that carbon intensity, the amount of carbon
dioxide released per joule of energy generated, has decreased slightly over
the past few decades as the world shifts from coal to cleaner natural gas.
Nevertheless, coal plants continue to be built, particularly in developing
countries such as China, and this prevents more rapid decarbonisation of
our economy.
Continued reductions in carbon intensity can be expected given the
flood of cheap natural gas that has arisen from the development and
application of new drilling techniques, in particular hydraulic fracturing
(more commonly known as fracking). Increases in renewables (e.g., wind
and solar) are also expected to reduce carbon intensity. Depending on
government policies, however, adoption of renewables could be slow or
rapid, leading to minimal or large reductions in carbon intensity. These
trends are expected to continue the long-term decline in coal usage,
although coal consumption is still increasing in certain places, such as
China.
Overall, increases in population (P) and affluence (A) have increased
faster than greenhouse-gas intensity (T) has declined, leading to an
increase in emissions of 75 percent between 1970 and 2005. Whether this
happens in the future depends in large part on how the world’s economy
evolves and what the world decides to do about climate change, as we
discuss in the next section.
281

8.3 Emissions scenarios
“It’s hard to make predictions – especially about the future?”1
Although we have a good idea of the factors that control greenhouse-
gas emissions, making accurate predictions of these factors is difficult. For
example, predicting future population trends requires predictions of factors
such as the rate of poverty, evolution of religious and social views on birth
control, the rate of education of women in high-fertility regions, available
healthcare in these regions, and so on.
Because of this difficulty, it is impractical to make a single prediction
of future emissions. Instead, the community of experts has developed a set
of alternative emissions scenarios. Each scenario is an internally consistent
vision of one way the world might evolve in the future, and the full set of
emissions scenarios is designed to span a plausible range of alternative
futures.
The scenarios most recently used by the IPCC in its predictions of
future climate change are known as the Representative Concentration
Pathways, frequently abbreviated RCP. The individual RCP scenarios are
named RCPx, where x is the radiative forcing in 2100. Thus, the RCP8.5
scenario has radiative forcing of 8.5 W/m2 in 2100, while the RCP2.6
scenario has radiative forcing of 2.6 W/m2 in 2100.
Each RCP scenario is associated with an internally consistent set of
assumptions for population, affluence, and technology. For example,
because people have fewer children as they get richer, the scenarios in
which the world’s poor become richer feature slower population growth
than the scenarios in which poverty is rampant. And the development and
adoption of new technology requires high economic growth to support it –
282

so the higher the economic growth scenarios also have more rapid
adoption of new and cleaner technologies.
Given an emissions scenario, the atmospheric concentrations of
carbon dioxide can be calculated by feeding those emissions into a carbon-
cycle model. The carbon-cycle model calculates how much of the carbon
dioxide emitted to the atmosphere is absorbed by the ocean and land
reservoirs. The remainder stays in the atmosphere and increases
atmospheric carbon dioxide.
Figure 8.2 shows yearly carbon dioxide emissions during the twenty-
first century. The RCP8.5 scenario is the most pessimistic – it assumes that
humans make essentially no effort to reduce emissions. As a result,
emissions increase throughout the twenty-first century, reaching 30
GtC/year by 2100 – about triple today’s emissions. Emissions level off and
decrease after 2150, reaching near-zero emissions by 2250.
Figure 8.2 Emissions of carbon dioxide for the four emissions
scenarios.
The emissions are in GtC per year (adapted from figure 6 of van
Vuuren et al., 2011).
Figure 8.3 shows that atmospheric carbon dioxide abundances
283

increase rapidly in response to RCP8.5’s huge emissions, with mixing
ratios exceeding 1,800 ppm in 2200 – this is more than six times the
preindustrial abundance of 280 ppm.
Figure 8.3 Atmospheric abundances of equivalent carbon dioxide (in
ppm) for the four emissions scenarios. Equivalent CO2 is the amount of
CO2 that gives the same radiative forcing as the full suite of radiative
forcers in the atmosphere in any year
(data downloaded from the RCP database: www.iiasa.ac.at/web-
apps/tnt/RcpDb).
The RCP6 and RCP4.5 scenarios assume that the world makes some
effort to reduce emissions; as a result, they have emissions peaking in the
middle of the twenty-first century. This leads to atmospheric carbon
dioxide stabilizing early in the next century at 750 ppmv and 540 ppm,
respectively – corresponding to about three times and twice preindustrial
carbon dioxide abundance.
The RCP2.6 scenario is the most optimistic scenario, with emissions
peaking around 2020 and then decreasing throughout the rest of the
century. After 2080, emissions actually become net negative, meaning that
carbon removal from the atmosphere (by, for example, growing plants and
284

http://www.iiasa.ac.at/web-apps/tnt/RcpDb

then burying the carbon) exceeds carbon released to the atmosphere. This
leads to atmospheric carbon dioxide peaking in 2050 at 440 ppm and
decreasing thereafter. By 2150 it is below present day values of 400 ppm,
and by 2500, it is 327 ppm – almost back to preindustrial carbon dioxide.
Achieving anything close to the RCP2.6 trajectory would require truly
heroic efforts.
285

8.4 Predictions of future radiative
forcing
Given the atmospheric abundances of carbon dioxide in Figure 8.3, along
with abundances of other greenhouse gases and aerosols, the radiative
forcing can be calculated. Figure 8.4a shows the radiative forcing
predicted for each scenario over the next 250 years. Forcing in the RCP8.5
scenario increases until about 2200, when it reaches more than 12 W/m2.
Forcing in the RCP6 and RCP4.5 scenarios stabilizes around 2100 at 6 and
4.5 W/m2, respectively. Radiative forcing from the optimistic RCP2.6
scenario declines continuously from its peak around 2020. By 2150 in this
scenario, radiative forcing has declined below present day values.
286

Figure 8.4 (a) Radiative forcing for the emissions scenarios (historical
radiative forcing is plotted before 2010). (b) Radiative forcing for the
RCP6 scenario, along with the radiative forcing in that scenario just
from carbon dioxide and from everything other than carbon dioxide
(data downloaded from the RCP database: www.iiasa.ac.at/web-
apps/tnt/RcpDb).
Figure 8.4b shows radiative forcing from the RCP6 scenario, along
with the radative forcing broken down into the contribution from carbon
dioxide and from everything else. The plot shows that, as we go into the
future, carbon dioxide is responsible for virtually all of the increase in
radiative forcing – the non-carbon dioxide component does not change
much. The reason for this is carbon dioxide’s long lifetime in the
atmosphere (discussed in Chapter 5) – once emitted, carbon dioxide stays
in the atmosphere for centuries. So carbon dioxide accumulates in the
atmosphere like water in a stopped up sink, and the radiative forcing from
it accumulates too. Other greenhouse gases, like methane, have a much
shorter lifetime (methane’s is about ten years), so it does not accumulate in
the same way that carbon dioxide does. This explains why there is such a
strong focus on carbon dioxide in policy debates over climate change.
287

http://www.iiasa.ac.at/web-apps/tnt/RcpDb

8.5 Predictions of future climate
288

8.5.1 Over the next century
The estimates of atmospheric radiative forcing shown in Figure 8.4 are
then input to climate models, which calculate a future climate for each
scenario. These are plotted in Figure 8.5 and they show that the set of
emissions trajectories translates into a wide range of future climates. The
large emissions associated with the RCP8.5 scenario lead to temperature
increases of 4°C over the twenty-first century, while the low emissions
associated with the RCP2.6 scenario lead to temperature increases of only
1°C, mostly during the first half of the century. If you want to relate the
warming to preindustrial temperatures, add about 0.8°C to each of these
numbers.
Figure 8.5 (left) Model estimates of global annual average surface
temperature for the four emissions scenarios (relative to the 1986–2005
average). The bars at right are the likely temperature difference between
the 1986–2005 period and the 2081–2100 period.
Adapted from figure SPM.7 of IPCC [2013].
Given the present political environment, the RCP2.6 scenario, which
has emissions peaking around 2020, appears hopelessly optimistic. This
RCP4.5 scenario, with emissions peaking around 2040 and atmospheric
carbon dioxide stabilizing around 540 ppm, seems (to me, at least) the best
289

we can hope for. That scenario yields warming of 1.8°C over the twenty-
first century. Of course, we might do worse than this, and even the RCP8.5
scenario is not out of the question. Given this, a reasonable estimate of
temperature increases over the twenty-first century might be 1.8–4.2°C.
It is worth noting that, despite huge differences in emissions, the
scenarios predict relatively similar warming until they begin to diverge
around 2040–2050. This occurs for two main reasons. First, emissions
reductions require us to fundamentally rebuild our energy infrastructure.
Doing this at any kind of reasonable cost means that it will take place over
several decades. Thus, even the most stringent efforts to cut emissions will
have only modest near-term impact (this can also be seen in emissions
estimates in Figure 8.2). Second, the high heat capacity of the ocean means
that any difference in radiative forcing needs to act for several decades
before significant differences in surface temperatures are evident.
The upshot is that the temperature trajectory over the next few
decades has already been determined largely by greenhouse-gas emissions
that have already occurred, investments in energy infrastructure that we
have already made, and the slow response of the climate system due to
thermal inertia from the ocean.
But Figure 8.5 also clearly shows that we do have significant control
over the amount of warming experienced by the end of the twenty-first
century. This is one of the many aspects of climate change that make it
difficult to solve. Addressing climate change will require us make
investments now and in the next few decades in renewable energy and
other energy efficiency technologies. But these investments will really
only pay off in the second half of the century. In other words, addressing
climate change requires people today to take actions that mainly benefit
future generations.
290

The right-hand panel in Figure 8.5 shows the likely range of
temperatures at the end of the twenty-first century predicted for all four
RCP scenarios. This range is generated by taking the same emissions
scenarios and running them through a large number of different climate
models. Because each climate model handles the details of the physics of
the climate system differently, the models produce slightly different
results. Thus, predictions of climate at the end of the twenty-first century
are uncertain because of uncertainty in which emissions pathway the world
will follow and also because of uncertainties in the physics of the climate
models.
An aside: Will the evolution of the climate over the twenty-first
century look like the trajectories plotted in Figure 8.5?
Not really. If you look at Figure 2.2a, you will see that, over the
past 130 years, the climate has generally warmed, but it also shows
significant year-to-year variability, which is caused by things such
as El Niño cycles and volcanic eruptions. Such variability means
that temperatures can decline for a few years, even as the climate is
experiencing a long-term warming (this was shown in Figure 2.5).
Individual model simulations also show this year-to-year
variability. The model lines plotted in Figure 8.5, however, are not
the result of individual model runs. Instead, each line is the
average of many model runs. In the individual model runs going
into the average, the highs and lows caused by the short-term
variability do not occur at the same time, so when you average
many model runs together, the short-term ups and downs tend to
cancel out and you get a smooth increase in temperature
throughout the century. In reality, short-term variability is going to
291

be important and we can expect the same kinds of ups and downs
seen in the past 130 years to continue to occur in the future.
292

8.5.2 Climate change beyond 2100
Even though Figure 8.5 stops in Year 2100, climate change does not stop
at that date. Exactly how long emissions can continue is a fiercely debated
point. Fossil fuels must eventually run out, and emissions from their
combustion will then cease. The range of total possible emissions until that
occurs extends from lower values of 1,500 GtC to more worrying
estimates of 5,000 GtC. These estimates are all well above the 370 GtC
that humans have already emitted into the atmosphere over the past few
centuries, and even the lowest estimates of carbon reserves would, if we
burned it all, lead to a manyfold increase in atmospheric carbon. It seems
likely to me that greenhouse-gas emissions will eventually cease because
of concern about climate change or because technological developments
make fossil fuels obsolete. But, given the deadlock in the public climate
change debate, it is difficult to predict when that will occur.
In Chapter 5, we talked about the long lifetime of carbon dioxide in
our atmosphere: Of carbon dioxide emitted into the atmosphere today,
about 25 percent will still be in the atmosphere in several centuries, and it
will take hundreds of thousands of years to remove all of the added carbon
from the atmosphere (Figure 5.9). The impact of the long residence time of
carbon dioxide in our atmosphere is shown in Figure 8.6a, which shows
atmospheric carbon dioxide over the next 1,000 years for emissions
scenarios in which atmospheric carbon dioxide increases until it reaches
550, 850, and 1,200 ppm, at which point emissions from human activity
decline instantly to zero.
293

Figure 8.6 (a) Amount of carbon dioxide in the atmosphere as a
function of time, for the next 1,000 years. Carbon dioxide emissions rise
at 2 percent/year until it hits a peak abundance (550,850, and 1,200
ppm); then emissions are decreased instantly to zero. (b) The
temperature time series corresponding to each carbon dioxide time
series
(adapted from Solomon et al., 2009, figure 1).
Even by the year 3000, eight to nine centuries after carbon dioxide
emissions ceased, atmospheric carbon dioxide in all scenarios remains well
above preindustrial values (280 ppm). This is simply a reflection of how
long it takes for an addition to atmospheric carbon dioxide to be removed.
The long-term evolution of temperatures associated with these carbon
dioxide time series are shown in Figure 8.6b. Even after emissions cease,
the temperatures do not significantly decline over the next 1,000 years.
This is a consequence of three factors. First, carbon dioxide remains
elevated throughout the millennium, so it continues to trap heat for a very
294

long time after emissions stop. Second, the ocean’s large heat capacity
means that the planet cools off very slowly because of its large heat
capacity. This is the flip side of the situation in which the warming lags the
carbon dioxide increase – the cooling will lag any decrease in atmospheric
carbon dioxide abundance. Third, slow feedbacks, such as the very slow
destruction of the planet’s big ice sheets, will act to oppose any cooling.
The important point here is that emitting large amounts of carbon
dioxide to the atmosphere this century commits the planet to elevated
temperatures for thousands of years. Once the temperatures rise, reducing
emissions will not bring the temperature back down quickly. We can
therefore think of climate change as being irreversible over any time
period that we conceivably care about. This also means that actions we
take today (or do not take) to curb emissions over this century will
essentially determine the climate for thousands of years. It is sobering
indeed to realize that people of the year 3000 or 4000 will be so affected
by actions we take today.
The irreversibility of carbon dioxide emissions can be usefully
contrasted with the second most important greenhouse gas, methane.
Methane has an atmospheric lifetime of ten years, meaning that a few
decades after emission, just about all of the methane is gone from the
atmosphere. Thus, if humans ever stop emitting methane, we would be
back down to its preindustrial value in a few decades.
295

8.6 Is the climate predictable?
One criticism of climate predictions goes something like this: “We cannot
predict the weather next week, so why does anyone believe predictions of
the climate in a hundred years?” This may sound reasonable, because it is
based on the correct observation that weather predictions are only accurate
a week or so into the future. However, the argument is built on a fatal flaw
– it makes the mistake of equating weather predictions with climate
predictions. In fact, it is possible to predict the climate in 100 years even if
weather is only predictable for a few days.
The root cause of this conundrum is that predicting the weather and
predicting the climate are fundamentally different problems. A weather
forecast is a prediction of the exact state of the atmosphere at an exact
time: “At 8 AM tomorrow, the temperature in Washington, DC, will be
3°C, and it will be raining.” If you get the time of an event wrong – for
example, you predict rain for 8 AM but it does not rain until 6 PM – then
you have blown the forecast. If you predict rain for the Washington, DC,
area but the rain falls 50 km to the west in Northern Virginia, then you
have blown the forecast. And if your temperature is off by a few degrees
and snow falls instead of rain, and it completely snarls traffic on Interstate
495, then you have really blown the forecast.
A climate prediction, in contrast, does not require predicting the exact
state of the atmosphere at any particular time; instead, it requires
predicting the statistics of the weather over time periods of years. Thus, a
climate prediction for the month of March for the years between 2080 and
2090 for a particular location might be as follows: average monthly
296

temperature of 12°C, with an average high of 16°C and an average low of
5°C; monthly average precipitation of 6.0 cm; and so on.
Being unable to make a prediction of the exact state of a complex
system (e.g., the weather) does not preclude the ability to predict the
statistics of the system (e.g., the climate). As an analogy, consider that it is
virtually impossible to predict the outcome of a single flip of a coin.
However, the statistics of coin flips are trivial: If you flip a coin 100 times,
I can tell you that you will get approximately fifty heads and fifty tails. In
other words, the inability to accurately predict any single coin flip does not
preclude the ability to predict the long-term statistics of the coin.
To make this point more concretely, answer the following question:
“Is it going to be hotter in Texas next January or next August?” If you
know Texas weather, you can predict with 100 percent certainty that
August is the hotter month, and you can make this prediction months,
years, or decades in advance. Think about that for a minute: You just made
a climate prediction that is valid years in advance – far beyond the ability
to predict weather.
More technically, weather forecasts belong to a class of problems
known as initial value problems. This means that, to make a good
prediction of the future state of the system, you must know the state of the
system now. If you have a marble rolling down a slope, and you want to
predict where it will be in one second, you need to know where it is now to
make that prediction. Similarly, to make a good weather forecast for
tomorrow, you have to accurately know the state of the atmosphere today.
The state of today’s atmosphere is then put into a forecast model, which
turns out a prediction of tomorrow’s atmosphere. However, small errors in
our knowledge of today’s atmosphere grow exponentially, so that a
forecast more than a few days in the future is dominated by the errors in
297

our knowledge of today’s atmosphere. That is why weather forecasts break
down after a few days.
Climate forecasts are a class of problems known as boundary value
problems. This type of problem does not require knowledge of today’s
atmospheric state but rather requires a knowledge of the radiative forcing
of the climate. This is why, for example, we can predict with 100 percent
certainty that August in Texas will be on average hotter than January in
Texas. We know this because we know that more sunlight falls on Texas
and the rest of the northern hemisphere during summer, leading to higher
temperatures.
Increases in greenhouse gases also increase the heating of the surface,
although by infrared radiation rather than visible. Thus, we can have
confidence that, if we add greenhouse gas to the atmosphere, the increase
in surface heating will warm the planet – just as we can predict that
summer will be hotter than winter.
One should not take this to mean that predicting the climate is an
easier problem than predicting the weather, only that they are different
problems. Some aspects of the climate problem are, in fact, harder than the
weather problem. For example, because weather forecasts cover only a few
days, weather models can assume that the world’s oceans and ice fields do
not change. Climate models, however, cannot make this assumption,
because both the world’s oceans and its ice fields can significantly change
over a century. Climate models must therefore predict changes in these and
other factors in order to accurately predict the evolution of the climate
system over a century.
298

8.7 Chapter summary
Prediction of future climate requires predictions of future emissions
of greenhouse gases from human activities. Such predictions are
known as emissions scenarios.
The factors that control emissions are population (P), affluence (A),
and greenhouse-gas intensity (T). This is expressed by what is
known as the IPAT relation: I = P × A × T, where I is carbon
dioxide emissions.
Greenhouse-gas intensity is the product of energy intensity and
carbon intensity. Energy intensity reflects the efficiency with
which the society uses energy as well as the mix of economic
activities in the society, with units of Joules of energy consumed
per dollar of economic output. The carbon intensity reflects the
technologies the society uses to generate energy, and it has units of
carbon dioxide emitted per Joule of energy produced.
Because predictions of the future are so uncertain, scientists have
constructed a set of plausible, alternative scenarios of how the
world might evolve. Taken as a group, these Representative
Concentration Pathways span the likely range of future emissions
trajectories.
Putting these emissions scenarios into a climate model yields
predictions of warming over the twenty-first century of 1.8 to
4.2°C. This is much larger than the warming of 0.8°C that the Earth
experienced over the course of the twentieth century.
299

Climate change does not stop in Year 2100. Carbon dioxide stays
in the atmosphere for centuries after it is emitted, so large
emissions of carbon dioxide this century will cause the Earth’s
temperatures to remain elevated for thousands of years.
Even though weather is not predictable beyond a few days, we can
nevertheless make climate predictions decades in advance. A
climate prediction is a prediction of the statistics of the system. For
many complex systems, predicting the statistics is easier than
predicting the specific state of the system.
300

Additional reading
Chapter 11 (Kirtman et al., 2013) and chapter 12 (Collins et al., 2013) of
the IPCC’s 2013 report detail, respectively, short-term and long-term
predictions of climate change.
Section 12.3 of Collins et al. (2013) describes the RCP scenarios. For a
more readable description, see SkepticalScience:
www.skepticalscience.com/rcp.php
D. Archer, The Long Thaw: How Humans Are Changing the Next 100,000
Years of Earth’s Climate (Princeton, NJ: Princeton University Press,
2010). Among the many things covered in this book is the very long-term
evolution of climate change.
See www.andrewdessler.com/chapter8 for additional resources for
this chapter.
301

http://www.skepticalscience.com/rcp.php

http://www.andrewdessler.com/chapter8

Terms
Carbon intensity
Emissions scenario
Energy intensity
Greenhouse-gas intensity
Gross domestic product (GDP)
IPAT relation
Representative Concentration Pathways
302

Problems
1.
a) Someone asks you about how much the climate will warm over
the next 100 years if we do nothing to address climate change.
How do you answer?
b) If the amount of carbon dioxide and other greenhouse gases
stopped increasing today and were held constant into the future,
how do you think the climate would change over the next century?
2.
a) Define each term in the IPAT identity.
b) What are the units of each term? Show how the units cancel so
that the I term has units of emissions of greenhouse gases.
3.
a) The T term can be broken into two terms. What are these two
terms, and what are their units?
b) If we switch from fossil fuels to solar energy, which of the terms
changes, and does this term increase or decrease?
c) If we convert from traditional incandescent lighting to LED
lights, which of the terms changes, and does this term increase or
decrease?
d) If we switch from natural gas to coal, which of the terms
changes, and does this term increase or decrease?
303

4. Consider this argument: “We cannot predict the weather in a week,
so there is no way we can believe a climate forecast in 100 years.” Is
this argument right or wrong? Explain your answer.
5. If we emit significant amounts of carbon dioxide this century, how
long will the planet remain warm?
6. Assume population grows at 2 percent/year and affluence grows at
3 percent/year.
a) How fast does the technology term have to decrease so that total
emissions do not change?
b) How fast does the technology term have to decrease to reduce
emissions by 20 percent in twenty years?
7. Explain how your level of wealth impacts how much emission of
carbon dioxide you are responsible for.
8. In 2002, the Bush Administration set a goal of reducing greenhouse
gas intensity by 18 percent by Year 2012. How ambitious is this goal?
What does this goal tell us about changes in emissions?
1 This statement has been attributed to various people, including Niels
Bohr and Yogi Berra.
304

9
Impacts of climate change
◈
Before the summer of 2010:
Russia is a northern country and if temperatures get warmer, it’s not
that bad. We could spend less on warm coats.
– Vladimir Putin, President of Russia1
After the summer of 2010:
Practically everything is burning. The weather is anomalously hot.
…What’s happening with the planet’s climate right now needs to be a
wake-up call to all of us, meaning all heads of state, all heads of
social organizations, in order to take a more energetic approach to
countering the global changes to the climate.
– Dmitri Medvedev, President of Russia2
You might have wondered as you read the first eight chapters, “OK, the
climate is changing because of human activities. Why should I care?” In
fact, warmer temperatures might sound good – you might associate them
with fun things like vacations at the beach or summer cookouts. But reality
is quite different. We rely in very important ways on the stability of the
305

climate for things such as food and fresh water. Most people do not notice
this reliance because it has been obscured by two centuries of scientific,
technological, and economic advancements.
Nevertheless, it is there. And every once in a while, an event comes
along that reminds us of the impact of climate on our lives. The heat wave
of the summer of 2010 was one of those events for the Russians. The
Russians learned the hard way that warmer temperatures do not mean tank
tops and grilled hot dogs but instead mean wildfires, loss of agricultural
crops, and suffering.
In this chapter, I cover the impacts of a changing climate. By the end,
I hope that you recognize that climate change is a significant risk that we
ignore at our peril.
306

9.1 Why should you care about climate
change?
In Chapter 8, we saw that if the world does nothing to address climate
change, we can expect global average temperatures to increase by a few
degrees Celsius during the twenty-first century. This may not seem like
much warming to you. After all, in many places summer days are 50°C
warmer than the winter days and daytime can be 25°C warmer than the
following night. And one day can be several tens of degrees Celsius
warmer or cooler than the next. If you consider the size of these
temperature variations, a change in the global average of a few degrees
may sound insignificant.
In this case, however, your intuition is wrong. Although the
temperature in any single place can vary considerably by season, by day,
and even within a day, the variations tend to cancel when averaged over
the entire globe. When you are experiencing the warmth of daytime,
someone on the other side of the globe is experiencing the coolness of
night. When it is summer where you live, it is winter in the other
hemisphere. Heat waves in one location are generally canceled by a cold
spell somewhere else (e.g., Figure 2.1). In other words, the large
temperature variations you experience are nearly completely canceled by
opposite variations somewhere else on the Earth.
Because of this cancellation, the global average temperature of the
Earth is very stable, with Figure 2.2a showing year-to-year temperature
variations of a few tenths of a degree. Moreover, seemingly small changes
in global average temperature are associated with significant shifts in the
Earth’s climate. For example, the global annual average temperature during
307

the last ice age was about 5°C colder than that of our present climate. At
that time, the Earth was basically a different planet: Glaciers covered much
of North America and Europe, leading to a very different distribution of
ecosystems, and because so much water was tied up in glaciers, sea level
was approximately 120 m lower than it is today.
And during the summer of 2003, a heat wave struck Europe in which
the average temperature in Europe was 3.5°C above average. Despite this
seemingly small amount of warming, this heat wave caused the deaths of
several tens of thousands of people. And temperatures today are perhaps
1°C warmer than they were a few hundred years ago, a period whose
climate was different enough that it has been dubbed the Little Ice Age.
Thus, we should take projections of a few degrees of warming seriously.
Furthermore, it is not just the size of the warming but the rate of
warming that is of concern. It took more than 10,000 years for the planet to
warm 5°C and emerge from the last ice age – an average rate of 0.05°C per
century. The rate of warming predicted for the twenty-first century is a few
degrees per century – about 100 times faster. Rate matters because the
faster the warming occurs, the less time people and natural ecosystems
have to adapt to the changes. If the sea level rises 1 m in 1,000 years, it
seems likely that we could adapt gracefully to that change. But a 1-m
increase in sea level in a century would be much harder to adjust to. And a
1-m increase in a decade would be a disaster, displacing millions of people
and destroying trillions of dollars of infrastructure.
Another argument often made is that a warming of a few degrees
should not cause concern because the Earth has gone through such
warmings and coolings many times during its 4-billion-year history. This
is undoubtedly true, as was discussed in Section 2.2. However, modern
human society, with a population of several billion people, metropolitan
areas with tens of millions of people, and reliance on industrial farming
308

and large-scale built infrastructure is only a century or so old. Over this
time, the Earth’s climate has been stable, varying by less than 1°C. Human
society as we know it has never had to face several degrees of warming in
a century. Thus, the argument that we have experienced this type of
warming before is fundamentally misleading.
Finally, you might be asking, “How do I know that a warmer climate
will not be better?” The reason it will not be is because both human society
and natural ecosystems have adapted to our present climate. If the climate
changes, then we will overall be less well adapted to our environment. As
an analogy, imagine that you go to the tailor and get a suit fitted exactly to
the shape of your body. At that point, no change in your body shape will
improve the way the suit fits – for example, either gaining or losing weight
will cause the suit to fit less well. In a similar fashion, any changes in the
climate, either warming or cooling, will result in overall negative
outcomes for human society.
As an example of how we rely on the climate, many structures in
Alaska are built on permafrost (ground that remains frozen year round)
with the implicit assumption that the ground will remain frozen. As long as
that is true, the structures are stable. If the permafrost melts, however, the
ground softens and can shift, potentially destroying structures (e.g.,
houses, bridges, roads) built on it. Thus, building on permafrost is a classic
example of relying on an unchanging climate. Unfortunately, given the
warming in the Arctic over the last century, assuming permafrost remains
frozen is turning out to be a bad bet.
Agriculture provides another example of adaptation to our present
climate. We farm where the climate provides suitable growing-season
temperature and precipitation. Around these farms we build essential
infrastructure to support agriculture: grain silos, processing plants, tractor
dealers, seed suppliers, and cities for all of these people to live. If the
309

climate shifts and the temperature and precipitation are no longer
conducive to farming, all of these investments to support agriculture will
no longer be useful. We may have to abandon them and rebuild the
infrastructure in whatever region becomes conducive to farming.
As a final example, imagine that you build a marina on the edge of a
lake. As soon as you pour the foundation, you are optimized for that
particular lake level. If the lake level goes up, your marina floods. If the
level of the lake drops, then you also face problems. If the drop is not too
much, you might be able to adapt with small changes like lengthening the
pier. But if the lake drops enough, then you must give up the original site
and rebuild the marina closer to the new edge of the water. In the Western
United States, where water levels are dropping in response to warmer
temperatures, increased demand, and drought, this process of following the
receding edge of the lake even has a name: “chasing water.”
Not every single change in every region will be negative. Reductions
in extreme cold events will have some benefits: less cold-weather
mortality, benefits to agriculture of fewer freezing events (which can
destroy some crops). Plant growth may well be enhanced in some regions
due to higher carbon dioxide abundances and increased water availability.
But these positive effects are expected to be outweighed by the more
pervasive negative effects.
The upshot of this discussion is that, when it comes to climate,
change is bad.
In this chapter, I will break the climate impacts into two components. I
will first discuss the physical impacts on the climate system: how
temperature, precipitation, sea level, extreme events, and other such
phenomena will change. I will then discuss the impact of these changes on
310

humans and those aspects of the environment that we rely on and care
about.
311

9.2 Physical impacts
312

9.2.1 Temperature
Although the global average temperature is currently increasing and will
almost certainly continue warming from each decade to the next, the
increase is not uniform across the globe. Figure 9.1 shows the distribution
of warming predicted for the various RCP scenarios, and there are several
key features in the warming distribution. First, continents warm more than
oceans because of the larger heat capacity of the oceans. As a result,
warming in northern North America and Eurasia is projected to be more
than 40 percent greater than the global average warming. Second, high
latitudes will warm more than the tropics. This is primarily due to the ice-
albedo feedback: The warming causes loss of ice, and the loss of ice
exposes dark ocean, which absorbs more sunlight and leads to further
warming (Figure 6.8). The models also predict more warming in the Arctic
than in the Antarctic. These changes are all continuations of trends that
have been observed over the past century (e.g., Figure 2.2b).
Figure 9.1 The distribution of annual-average warming in the middle of
313

the twenty-first century (left-hand column), the end of the twenty-first
century (center column), and the end of the twenty-second century for
the four RCP scenarios. Temperature increases are relative to the
1986–2005 average. These are calculated from an ensemble of climate
models, with the number of models indicated in the upper right corner
of each panel. Hatching indicates regions where there is disagreement
among the models on the sign of the change, while stippling indicates
regions where the models all agree on the sign of the change.
This figure is adapted from figure 12.11 in Collins et al. [2013].
In general, adding greenhouse gases to the atmosphere tends to
reduce temperature contrasts. We see this in the enhanced warming in the
Arctic, which tends to reduce the temperature difference between the
tropics and polar regions. We also expect more warming in winter than in
summer (reducing summer-winter temperature contrasts), and more
warming at night than during the day (reducing day-night temperature
contrasts). To understand why this happens, you need to recognize that
temperature variations in our climate are generally caused by variations in
the distribution of sunlight. The polar regions, nighttime, and winter are all
colder because they receive less sunlight than the tropics, daytime, and
summer. The atmosphere also heats the surface, but because greenhouse
gases tend to be well mixed in our atmosphere (because of their long
residence times), the heating from greenhouse gases occurs evenly over
the entire surface of the planet – it is the same at night as during the day,
the same in winter as in summer, the same at high and low latitudes.
As the abundance of greenhouse gases increases, heating of the
surface from greenhouse gases becomes stronger while heating from
sunlight remains about the same. Thus, variations in solar heating with
latitude, time of day, and season become a smaller component of the total
314

heating of the surface, and will therefore lead to smaller temperature
variations. In the limit of a planet such as Venus, with a massive
greenhouse-gas-rich atmosphere, the heating of the surface by the
atmosphere is roughly 16,000 W/m2, which is about eighty times larger
than solar heating. As a result, the variations in the solar input are so small
in comparison that the temperature everywhere on Venus is approximately
the same: 735 K.
Figure 9.1 shows that, in the middle of the twenty-first century, all of
the RCP scenarios show about the same amount of warming. This small
difference reflects the large inertia in our climate system from the large
heat capacity of the oceans, as well as inertia in our economy – any policy
to address climate change at reasonable cost will phase out fossil fuels
over a few decades (phasing them out immediately would be very, very
expensive). As a result, the warming over the next few decades is already
determined by past emissions and our present mix of energy technology.
By the end of the twenty-first century, however, the scenarios have
diverged: global average warming in the RCP8.5 scenario is four times
larger than in the RCP2.6 scenario and twice that found in the RCP4.5
scenario. And the differences are even larger at the end of the twenty-
second century. Thus, while our actions will not have much effect for the
next few decades, policies to address climate change allow us to avoid the
very large warming predicted for the second half of the twenty-first
century and beyond.
Overall, we can have high confidence in the general shape of these
predictions. There is, of course, uncertainty in the exact magnitude of the
warming, both from our understanding of the climate system and from
uncertainty in how emissions will change in the future. But we can be
confident that temperatures will continue to increase in the future and that
315

the general distribution of the warming will be in accord with these model
predictions.
316

9.2.2 Precipitation
As greenhouse gases increase, Ein for the surface increases because of
increased infrared radiation from the atmosphere falling on the surface.
This leads to an increase in evaporation from the oceans, and because
precipitation must balance evaporation, precipitation also increases. More
quantitatively, total global precipitation is projected to increase by a few
percent for every degree Celsius of global average warming.
Although total rainfall is expected to increase, the increase will not be
distributed evenly. Predictions of changes in annual average precipitation
from a set of climate models are shown in Figure 9.2. There is a large-
scale shift of precipitation to higher latitudes, causing a decrease in many
parts of the tropics. A general rule of thumb is that wet places will get
wetter, while dry places will get drier.
Figure 9.2 Change in annual mean precipitation over the twenty-first
century as predicted by climate models driven by the RCP8.5 high-
emissions scenario. The change is the percent change of 2081–2100
precipitation relative to the 1986–2005 precipitation. This is the average
of thirty-nine models’ predictions. Hatching indicates regions where
there is disagreement among the models on the sign of the change, while
317

stippling indicates regions where the models all agree on the sign of the
change.
(adapted from IPCC 2013, figure SPM.8).
In addition to changes in the amount of precipitation, there will also
be shifts in the form. Less wintertime precipitation will fall as snow and
more will fall as rain. This is more important than it might sound: When
snow falls, the water does not run off until the snow melts in spring. Rain,
on the other hand, runs off immediately, so changing the form of
precipitation will change the timing of runoff, which has important
implications for water availability.
These predictions of future precipitation patterns (like Figure 9.2) are
less certain than the temperature predictions (like Figure 9.1). Precipitating
clouds can be small (sometimes just a few kilometers across), and climate
model grids are too coarse to directly simulate such small structures. In the
end, we can have some confidence in the predictions at a qualitative level
(e.g., shifts in precipitation location, more rain falling in heavy events) but
confidence should be lower for the quantitative details of the spatial
distribution of the changes.
318

9.2.3 Sea-level rise and ocean acidification
Sea-level rise is one of the most certain impacts of climate change. As we
learned in Chapter 2, the sea rises in response to warming temperatures for
two reasons. First, as grounded ice melts, the melt water runs into the
ocean, increasing the total amount of water in the ocean and, therefore, sea
level. Second, like most things, water expands when it warms, which also
tends to raise sea level. Measurements (Figure 2.10) confirm that sea
levels have indeed been rising as temperatures have gone up, and we can
be certain that the seas will continue to rise into the next century.
The latest IPCC report predicts that sea level will rise 45 to 75 cm (18
to 30 inches) above today’s levels by 2100. Even the low end is nearly
three times the sea-level rise of the twentieth century of 17 cm (7 inches).
But even more worrying is that we have strong evidence from the last
interglacial (about 120,000 years ago), when temperatures were about 2°C
warmer than today, that sea levels were at least 5 m (16 ft) higher, mainly
due to melting of the Greenland ice sheet. Given that we are likely to see
this amount of warming above preindustrial temperatures during the
twenty-first century, why are the estimates of twenty-first-century sea-
level rise (relatively) small?
The reason is that it takes a very long time – many centuries, or even
millennia – to melt this much ice. This means that warming temperatures
this century and beyond will likely guarantee a few meters of sea-level
rise, but we will only see a small fraction of this this century. Most of the
increase in sea level due to our emissions will occur far in the future. Put
another way, our actions today might commit residents of the Earth in the
year 2500 or 3000 to live in a world with much higher seas.
319

Ocean acidification is another certain consequence of continued
emissions of carbon dioxide. As we explored in Chapter 5, a significant
fraction of carbon dioxide emitted to the atmosphere by humans ends up in
the oceans. In the liquid environment of the ocean, carbon dioxide reacts
with water and is converted into carbonic acid (Equation 5.3). The net
result is that, as the oceans absorb more and more carbon dioxide, the
oceans will become more and more acidic.3 In fact, since the industrial
revolution, the absorption of carbon has lowered the ocean’s pH by
approximately 0.1. At present, the ocean is now more acidic than it has
been for 20 million years.
320

9.2.4 Loss of ice
Ice melts reliably at 0°C, so we expect the planet’s ice to vanish as the
globe warms. Indeed, we have observed a steady retreat of the world’s
glaciers (Figure 2.6), Arctic sea ice (Figure 2.7), and the Greenland and
Antarctic ice sheets over the past few decades (Figure 2.8). This is one of
the main contributors to the observed increase in sea level.
In the future, we can expect sea ice to continue to decrease in the
Arctic, and to begin to decrease in the Antarctic as temperatures there
begin to increase. Most predictions are that the Arctic ocean will be
entirely ice free during summertime at some point in the twenty-first
century, although exactly when is still being debated.
321

9.2.5 Extreme events
Extreme weather events – heat waves, droughts, floods, hurricanes,
tornadoes, to name just a few – are some of the most consequential ways
that the environment affects us. If a warming climate causes increases in
the frequency or severity of these types of events, this would be a scary
and consequential outcome of the burning of fossil fuels.
Some changes in extremes are certain, or nearly so. We can be certain
that extreme heat events will become more frequent and severe. The
impacts of this are serious, as evidenced by the tens of thousands of people
who perished during the European heat wave in the summer of 2003. In
fact, extreme heat is the largest weather-related cause of death in the
United States.
Figure 9.3 shows a climate model calculation of how temperature
extremes might change in a warmer climate under a high-emissions
scenario (RCP8.5). The panel shows the distribution of daily average
August temperatures around Dallas, Texas, for the periods 2006–2015 and
2091–2100. The daily average temperature increases between these two
periods from 28°C to 36°C – an increase of 7.8°C (14°F). Figure 1.1
showed a similar, although much smaller, shift in the frequency
distribution of daily temperatures in Fairbanks, Alaska, in response to the
warming of the twentieth century.
322

Figure 9.3 The frequency of occurrence of daily average temperatures
around Dallas, Texas, in two periods: 2006–2015 and 2091–2100.
Temperatures come from the GFDL-CM3 climate model driven by the
high-emission RCP8.5 scenario.
We can estimate the occurrence of today’s extreme temperatures as
the temperature exceeded by the warmest 5 percent of days between 2006
and 2015. In the climate model, this corresponds to days with average
temperatures above 32.4°C. In the 2091–2100 period, temperatures above
32.4°C are projected to occur on 78 percent of the days – a dramatic
increase in extreme temperatures.
The hottest day of the 2006–2015 period had a daily average
temperature of 36.5°C. Disturbingly, 39 percent of the days between 2091
and 2100 will be hotter than this. This means that the climate is reaching
uncharted territory: temperatures that never occur today will be occurring
frequently by the end of this century. We can verify this trend in the
historical record, where we see record heat more often than record cold. In
2012, for example, there were 34,008 daily high temperature records in the
United States and 6,664 daily low temperature records.4
323

In addition to changes in the pattern of precipitation (Figure 9.2), it is
likely that more rainfall will come in the heaviest downpours, which
continues a trend observed over the last few decades. During a heavy
downpour, the soil saturates before the end of the rain event, and the
remaining rain therefore runs off, leading to a number of negative
consequences, such as increased risk of flooding and loss of freshwater for
use by humans and ecosystems.
An increase in the fraction of heavy events also tends to increase the
time between rain events. Combined with warmer temperatures, which
will increase the rate at which water is lost from soils by evaporation, this
will also contribute to increased drought occurrence. Thus, we get the
surprising result that both wet and dry extremes will grow more likely in
the future: wet extremes, with associated risks of flooding, increased
erosion, and landslides; and dry extremes, with associated risks of water
shortages and drought.
One topic that is frequently talked about is hurricanes, typhoons, and
cyclones (these storms are the same; the name is determined by which
ocean the storm appears in). These are some of the most dramatic weather
events we face, and they can cause enormous damage. So how will
hurricanes and their impacts change in the future? First, we can say with
certainty that the impacts of hurricanes will be more severe because a
rising sea level will lead to higher storm surges and flooding, as well as
increased rates of coastal erosion. As far as the storms themselves go,
research indicates that hurricanes may get stronger, although the number
of these storms may decrease. But our confidence in this impact is far from
certain.
324

9.3 Impacts of these changes
As described in the previous section, climate change will bring about a set
of certain impacts (e.g., increasing temperatures including extreme heat
events, changes in precipitation, increases in sea level, increasing acidity
of the ocean) as well as a large number of potential impacts (e.g.,
increasing flood and drought intensity and frequency, increases in
hurricane intensity). These physical changes in the climate system are only
the first step in determining climate impacts. Determining impacts also
requires an understanding of how vulnerable human and natural
ecosystems are to these changes.
One thing that makes this a challenging exercise is that humans can
adapt to a changing climate. For example, many buildings in Boston are
currently not air-conditioned, because there are only a few days a year that
are hot enough to require cooling. As the climate warms, air-conditioners
can be installed in buildings as needed to accommodate the heat. Similarly,
increases in sea level can be dealt with by building seawalls or by
relocating people. And many ecosystems that humans rely on are
intensively managed, such as agriculture, commercial forests, rangelands,
and fisheries. Because human management dominates these systems, the
ability to adapt management practices to changing conditions offers the
possibility of mitigating at least some of the harmful impacts. At the same
time, disruption of these systems by climate change may have severe
human impacts because we depend on them so much.
Agriculture is probably the most obvious economic sector that is
going to be impacted by climate change – through changes in temperature
(photosynthesis in some of our most important crops is most efficient at
325

temperatures between 20°C and 25°C) and fresh water availability (a good
rule of thumb is that it takes 1,000 tons of water to produce 1 ton of grain,
and about 15,000 tons to produce 1 ton of meat).5 Adjusting farming
practices can alleviate most of the negative impacts for local warming
below about 3°C. Above this threshold, food production is projected to
decrease. As Figure 9.1 shows, the continents warm faster than the ocean,
so 3°C local warming in agricultural areas might occur as early as the
middle of the twenty-first century in the high-emission scenarios. A
decrease in food production beginning in just a few decades would
constitute a massive challenge given the projected increase in population
during the twenty-first century.
This is just one example of how changes in the climate will affect us.
In particular, changes in fresh water will be a major challenge of climate
change because we rely on it for so many things. For example, most
electricity power plants require huge amounts of water for cooling, which
is why these plants are usually situated on rivers or lakes. During droughts,
low water levels sometimes force these plants to shut down.
In some regions, overall water availability is expected to increase. For
example, river runoff is projected to increase by 10 percent to 40 percent
by mid-century at higher latitudes and in some wet tropical areas,
including populous areas in East and Southeast Asia. However, the
beneficial impacts of increased annual runoff in these areas may be
tempered by changes in the timing of the runoff. For example,
summertime runoff from melting snowpack provides an important source
of fresh water to the U.S. Pacific Northwest at a time when there is little
rainfall. Warming temperatures, however, will lead to less wintertime
precipitation falling as snow and more as rain, and the snow that does fall
will melt earlier. Both of these effects will tend to shift runoff from the
summertime, when the water is most needed, toward winter and spring,
326

when it is less needed. This will increase stresses on summertime water
supplies.
In the mid-latitudes and dry tropics (e.g. the Mediterranean Basin,
Western United States, southern Africa, and northeastern Brazil),
decreases in rainfall and increases in temperature will lead to a significant
decrease in water resources. Drought-affected areas in these regions are
expected to increase in extent. Humans and human-managed ecosystems
should be capable of adapting to decreases in fresh water supplies, as long
as the decreases are modest. Large reductions in fresh water supplies
associated with the highest warming scenarios, however, would likely
exceed our ability to respond gracefully.
Other impacts of climate change will be much harder to manage.
Many of these are on systems that humans do not manage because they are
fundamentally unmanageable, such as rising sea level. As discussed
earlier, sea level is predicted to rise 45 to 75 cm (18 to 30 inches) above
today’s levels by 2100. This is a huge challenge for human society because
a significant fraction of the world’s population lives within a few feet of
sea level. Moreover, some of the world’s most productive farmland is
located in river deltas and other regions that are particularly sensitive to
sea-level rise.
Even small amounts of sea-level rise will therefore have significant
negative implications. In Florida, for example, a sea-level rise in the
middle of the projected range would inundate 9 percent of Florida’s current
land area at high tide.6 This includes virtually all of the Florida Keys as
well as 70 percent of Miami-Dade County. Almost one-tenth of Florida’s
current population, or nearly 2 million people, live in this vulnerable zone,
and it includes residential real estate now valued at over $130 billion. It
also includes important infrastructure, such as two nuclear reactors, three
prisons, and sixty-eight hospitals. And this is just Florida. Multiply these
327

impacts to account for all of the places on the planet where people live
near sea level, and you can get a feel for how big a problem this is going to
be.
And regions do not actually have to be submerged to be affected.
Increased sea level will increase the frequency of flooding from extreme
sea-level events, so that a flood event that occurred, say, every 100 years,
may occur every few years by the end of the century. As the flooding of
New Orleans after Hurricane Katrina showed, these events cause
significant loss of life as well as economic destruction.
It is worth noting that sea-level rise provides a particularly good
example of how we are adapted to our present climate. When cities such as
New Orleans or Miami were founded, the original inhabitants did not take
into account the possibility of sea-level rise. Rather, they assumed that the
sea level would remain pretty much as they found it, so they built their city
in concert with that particular sea level. Any change in sea level, either up
or down, has negative effects on these cities.
Humans can adapt to rising sea levels, of course, but all of the
adaptation solutions are hard and expensive. Our basic choice is between
building extremely expensive infrastructure to protect the city (e.g., sea
walls) or simply abandoning those areas and the trillions of dollars of
infrastructure in them. Given the certainty of sea-level rise, the most
rational approach would be to immediately cease building new
infrastructure in regions that are likely to be overrun by sea-level rise in
this century (i.e., Miami) and instead begin an orderly multi-decadal
retreat from the coast. That, of course, is not happening.
Another unmanageable impact of climate change is the acidification
of the ocean. Many ocean organisms build shells or skeletons out of
calcium carbonate, and their ability to do this is affected by the acidity of
the ocean. As the ocean becomes more acidic, these species will at first
328

find it more difficult to extract carbonate from the water for use in their
shells or skeletons. Eventually, the acidity will increase to the point where
it is fatal for the species. It is important to realize that ocean acidification
is not just a theory – it has happened before. During the PETM (discussed
in Section 7.5), a massive amount of carbon was emitted into the
atmosphere, which subsequently dissolved into the ocean. That event was
accompanied by an acidification of the ocean that dissolved much of the
carbonate sediment there (as shown in Figure 8.6).
Some species will adapt better to increasing acidity than others. As a
result, the mix of species in ocean ecosystems will shift, resulting in new
and novel arrangements of species. This will affect humans because we
rely in important ways on the ocean: e.g., about a billion people rely on the
ocean as their primary source of protein. And, like sea-level rise, this is
basically an unmanageable impact. If the amount of protein available for
human consumption from the ocean decreases, there are no simple
adaptations to solve that problem – the protein will have to be made up
elsewhere or people will starve.
Natural ecosystems and their constituents will also be affected. At the
individual species level, for example, research shows that warming
temperatures are presently driving lizards to extinction. During spring,
when energy demands are highest because lizards are reproducing, the
warming temperatures reduce the amount of time that lizards can forage
for food (if the temperatures get too high, cold-blooded lizards have to
rest). If temperatures continue to increase (which we expect), then at some
point the time available to look for food diminishes to the point where
lizards simply cannot find enough food – and extinction ensues. This is
already happening, and extrapolating into the future, global warming may
lead to the extinction of 40 percent of all global lizard populations by
2080. Note that it is not the global average temperature that matters to
329

lizards, or even the local average temperature, but the local daily
temperatures during one particular time of the year. This emphasizes that it
is the details of climate change that ultimately matter, not the broad-brush
changes in global average quantities.
Now you may not care much about lizards, but you should care about
their extinction for two reasons. First, the environment is a tightly coupled
system. There are many examples in history in which humans have
intentionally removed a species from the environment because they
thought it was harming them (e.g., getting rid of birds because the birds
were eating crops) only to find out that that change led to more problems
than it solved (e.g., the birds were also eating insects, and with the birds
gone the insects proliferated and destroyed much more of the crops than
the birds were eating). Today’s modern world obscures many of these
relationships, but they nonetheless still exist. Removing lizards from an
ecosystem may have important effects on the rest of the environment that
we do care about, just like pulling a single thread on a sweater can unravel
the entire thing.
Second, this is not just about lizards. As warming temperatures drive
lizards to extinction, the same warming temperatures will be having
deleterious effects on many other species. In fact, a significant fraction of
plant and animal species may be at increased risk of extinction if global
average temperatures increase by a few degrees Celsius. Thus, at the
expense of a mangled metaphor, lizards may be the canary in the coalmine.
Changes to individual species will project onto changes in entire
ecosystems, such as alpine meadows or temperate forests. As the climate
changes, each component species of an ecosystem will be affected in its
own way. Some species may adapt readily, whereas others may be unable
to adapt fast enough to survive. Species will also be subject to human
330

interventions and constraints such as land-use change, barriers, and
intentional or inadvertent transport.
The aggregate result will be that ecosystems will evolve, with new
relationships among incumbents and new arrivals developing in each
location. In some cases, the new assemblages may be similar enough to
present ecosystems that we can think of them as basically unchanged. In
other cases, however, the new systems may be unlike any present
ecosystems, with new species and relationships between them and other
ecological surprises. Some ecosystem types are likely to be lost entirely,
such as alpine systems, coastal mangrove systems, and coral reefs.
An aside: What are ecosystem services?
Natural ecosystems provide enormous benefits to human society.
The mangrove forests that grow in shallow salt-water coastal
regions are good examples. They provide important protection for
coastal areas from erosion, storm surge (especially during
hurricanes), and tsunamis. Their loss will cost us money – either
we will have to build expensive coastal defenses to replace the
natural defense provided by the mangrove or we will have to
absorb the cost of increased coastal damages. This value, provided
to humans for free, is what we mean when we talk about
ecosystem services.
Another good example is pollination by bees. Many crops
(e.g., apples, almonds, blueberries) are directly dependent on bee
pollination as part of their growing cycle, and the total value of
these pollinated crops in 2010 was $16 billion. The cost of
pollination (by wild bees, at least) is provided free of charge by
nature to us. In China, a decline of wild bees has forced farmers to
331

hire people to go from flower to flower and hand-pollinate the
flowers using tiny brushes. Thus, ecosystems provide important
economic benefits to our society; the impacts on them from climate
change will therefore impose potentially steep costs on society.
A particularly important factor in determining the severity of the
impacts on natural ecosystems will be the rate of climate change.
Ecosystems have adapted to large climate change in the past, such as the
warming from the previous ice age. However, the warming predicted for
the next century will be incredibly fast – perhaps one hundred times faster
than the average rate of warming since the last glacial maximum 20,000
years ago. As the rate of warming goes up, the ability of the environment
to gracefully adapt to the changes declines. What is uncertain is how much
less gracefully, and with what consequences.
We can also expect human health to be negatively impacted by
climate change. Some of these health impacts follow closely from changes
already discussed, such as negative health consequences of warmer
temperatures or malnutrition associated with reductions in food
availability. In addition to those impacts, we expect warmer, more humid
days to enhance the photochemical reactions that cause air pollution,
leading to more smoggy days as the climate warms, along with the
associated health impacts.
Warming temperatures also increase disease risk as a result of
expansions in ranges of animals that transmit the diseases (e.g.,
mosquitoes), shortening of the diseases’ incubation periods, lack of very
cold temperatures that can kill the transmitters, and disruption and
relocation of large human populations. Moreover, increases in water
temperature, precipitation frequency, and other factors could increase the
332

incidence of water contamination with harmful pathogens, resulting in
increased human exposure.
However, because human health is part of an intensively managed
human health care system, we expect human society to adapt and
effectively manage many of the impacts – for modest warming, at least. As
the warming gets progressively larger, the ability and effectiveness of
strategies to manage the health threats from climate change will decrease
and the net negative impact on human society will grow rapidly.
It should also be noted that intensively managing climate change
costs money. This means that the ability to adapt is not evenly spread
across the globe. Rich, well-governed places such as the United States or
Europe have resources that can be applied to adapting to climate change.
For small climate change (i.e., the bottom end of the range in Figure 8.5),
rich countries will likely find most effects of climate change to be
manageable without too much social disruption. If climate change falls
toward the upper end of the predicted range in Figure 8.5, then climate
change is expected to be a serious, perhaps insurmountable challenge for
even these rich countries.
Approximately 2 billion people, however, are so crushingly poor that
they have no additional resources available to address even minor climate
change. For them, installing air-conditioners or building coastal defenses
in response to rising seas, developing new freshwater infrastructure in
response to water shortages, improving public health infrastructure in
response to a new disease outbreak, and the like are simply not affordable
options. For these poorest of the poor, even small climate change will be a
major challenge. If climate change falls toward the upper end of the
predicted range in Figure 8.5, then it would be a certain disaster for the
poorest.
333

Because some countries will be unable to manage even moderate
climate change, many experts view climate change as a potentially
destabilizing social force. Wars over resources (e.g., fresh water) or large-
scale migration to escape climate extremes (e.g., droughts) could cause a
range of social and economic disruptions, leading to failed states and the
associated impacts on regional and world stability. This has led national
defense analysts at the Pentagon to categorize climate change as a threat to
U.S. national security.
334

9.4 Abrupt climate changes
Many of the changes I just described – changes in temperature,
precipitation patterns, sea level, and so on – are steady changes in the
climate system. For example, we expect the climate to warm by a few
tenths of a degree Celsius per decade (as suggested by plots such as Figure
8.5), and we expect sea level to rise by a few centimeters per decade.
These changes, while incredibly fast geologically, are gradual on human
timescales.
An abrupt climate change is a sudden and significant shift in some
aspect of the climate. As an analogy, imagine that you are sitting in a
canoe and you start to lean over. At first, the canoe tilts with you – until,
that is, you pass a critical threshold and the canoe suddenly flips over,
throwing you and everything else in the canoe into the river. That is an
abrupt change.
For the climate, the worry is that the climate will not warm smoothly
as greenhouse gases are added to it. Rather, we will add enough
greenhouse gas that the climate system will undergo a large and rapid shift
to an entirely new climate state – equivalent to the canoe rapidly
transitioning from right side up to upside down. In the case of climate, the
large climate shift might occur on a timescale of decades.
This possibility concerns scientists because abrupt changes have
happened in the past. During the PETM approximately 55 million years
ago, there was a rapid release of greenhouse gases and a subsequent
warming of 5–9°C in just a few thousand years. It is not known what
caused the release of greenhouse gases, but one possibility is that it was
due to a carbon cycle feedback – for example, an initial warming melts
335

permafrost, leading to the release of carbon stored in it, which leads to
more warming, and so on. Scientists worry that warming due to humans
could do the same thing: melt permafrost and release massive amounts of
carbon dioxide and methane. That would doom the planet to another
PETM.
In addition, roughly 12,000 years ago, as the Earth was emerging
from the depths of the last glacial maximum, the temperature suddenly
plunged (at least in the mid-and high latitudes of the northern hemisphere).
The period of low temperatures during the millennium that followed, today
known as the Younger Dryas, is thought to have been due to a massive
release of water into the North Atlantic from melting glaciers. This
freshwater influx disrupted the ocean currents, in particular the Gulf
Stream. Because the Gulf Stream transports heat from the tropics to the
high latitudes, the shutdown of the Gulf Stream caused mid- and high-
latitude temperatures to plummet (this was the basic scientific premise
behind the movie The Day After Tomorrow).
Thus, abrupt changes do happen and we must take their possibility
seriously. However, beyond acknowledging the possibility, there is little
the scientific community can say. Climate models do not predict the
occurrence of an abrupt climate change, and most experts view the
probability to be low, but not zero, over the coming century. If an abrupt
change did occur, though, it could be a catastrophe. Such low-risk, high-
consequence events pose significant challenges to our society. The
tendency is to ignore the risk until it occurs, which is why dams are built
after floods, and not before. However, for these types of events, once the
abrupt change takes place, it will be very difficult, if not impossible, to
gracefully manage. This makes the strategy of ignoring the risk a
precarious proposition.
336

It is also possible that abrupt impacts can occur in a slowly changing
climate when the climate reaches a threshold set by human vulnerability. A
good example is the flooding of lower Manhattan during Superstorm
Sandy. As the storm surge reached certain levels, abrupt and large impacts
were suddenly felt. For example, as long as the water level stayed below
the entrance to the subways, there was no flooding of the subway system.
But as soon as the water overtopped the entrance to one station, water
rushed into the system and, traveling through the subway tunnels, was able
to flood large parts of the system. Some sections were out of commission
for months, and the cost of fixing it was immense.
337

9.5 Chapter summary
The amount of warming predicted for the twenty-first century (a
few degrees Celsius) is comparable to the warming since the last
ice age (5°C). This means that the warming over the next century
or two may herald a literal remaking of the Earth’s environment
and our place within it.
We are adapted to our present climate, so any significant change
(in any direction) is likely to be detrimental.
There are a number of virtually certain impacts of climate change:
The climate will get warmer (with more extreme heat events),
precipitation patterns will change, sea level will rise, and the
oceans will become more acidic. These are serious impacts that
should compel our attention.
There are also a number of more speculative changes in extreme
events (e.g., floods, droughts, hurricanes). Many of these changes
are likely, but we cannot be sure of the severity.
These changes may have important negative impacts on humans.
This includes impacts on agriculture and freshwater availability, as
well as public health consequences.
Natural ecosystems may also be severely disrupted. These natural
systems provide services of great value to humans, so their
disruption could provide significant challenges to us.
The impacts of these changes on human society will not be
distributed evenly. The wealthy countries of the world will have an
338

easier time adjusting than will the poor countries of the world, and
there may be some people who benefit, particularly if the warming
is small. As the amount of warming increases, negative impacts
will increasingly dominate the benefits, and even the richest
countries will be severely challenged.
Abrupt changes are low-probability, high-consequence events. An
example of an abrupt change was the reorganization of the ocean’s
circulation during the Younger Dryas period about 12,000 years
ago. Although scientists do not expect them to occur this century,
they cannot be ruled out.
339

Additional reading
Much has been written about the impact of climate change on humans and
the rest of the environment. Here are a few notable works.
The IPCC’s Working Group II focuses on impacts of climate change and
our ability to adapt to them. You can download and read the IPCC’s Fifth
Assessment Report at ipcc-wg2.gov/AR5/report/.
The IPCC has also released an assessment of the links between climate
change and extreme weather and climate events, the impacts of such
events, and the strategies to manage the associated risks. See C. B. Field,
V. Barros, T. F. Stocker, D. Qin, D. J. Dokken, K. L. Ebi, M. D.
Mastrandrea, K. J. Mach, G.-K. Plattner, S. K. Allen, M. Tignor, and P. M.
Midgley (eds.), Managing the Risks of Extreme Events and Disasters to
Advance Climate Change Adaptation. A Special Report of Working
Groups I and II of the Intergovernmental Panel on Climate Change
(Cambridge and New York: Cambridge University Press, 2012), 582 pp.
(download at ipcc-wg2.gov/SREX/).
William Nordhaus, The Climate Casino: Risk, Uncertainty, and
Economics for a Warming World (New Haven, CT: Yale University Press,
2013). This is an excellent book about the economics of the climate
problem. Part II of the book contains a first-rate description of how
economists evaluate climate impacts.
Here are a few other books I recommend that describe the impacts of
climate change. They are all aimed at the general public, so they are easy
reads and require no specialized knowledge: E. Kolbert, Field Notes from
a Catastrophe: Man, Nature, and Climate Change (New York:
340

http://ipcc-wg2.gov/AR5/report/

http://ipcc-wg2.gov/SREX/

Bloomsbury USA), 2006; M. Lynas, Six Degrees: Our Future on a Hotter
Planet (Washington, DC: National Geographic, 2008); J. Diamond,
Collapse: How Societies Choose to Fail or Succeed (New York: Penguin,
2011).
See www.andrewdessler.com/chapter9 for additional resources for this
chapter.
341

http://www.andrewdessler.com/chapter9

Terms
Abrupt climate changes
Ecosystem services
342

Problems
1. Your third cousin once removed asks you why we will not be better
off in a warmer climate. What do you tell him?
2. Your friend says, “Climate scientists are such alarmists. First they
say that floods will become more frequent, and then they say that
droughts will become more frequent. Come on, which one is it? They
cannot both occur!” What do you tell her?
3. As discussed in this chapter, temperatures are not expected to rise
uniformly across the globe.
a) Why is there more warming at high latitudes than the tropics?
b) Why will land warm more than the ocean?
c) Why do temperature contrasts (e.g., night vs. day) decrease in a
warmer climate?
4. Precipitation
a) How is precipitation expected to change in a future climate?
b) Why do changes in the form of precipitation (rain vs. snow)
matter?
5. Explain a few ways that climate change impacts public health.
6. Why will it be easier for the United States and Western Europe to
deal with climate change than countries in Africa?
7. Explain how climate change affects our national security.
8.
343

a) What do scientists mean when they talk about “abrupt climate
change”?
b) Give an example of an abrupt climate change that has occurred
in the past.
1 Quoted in “Nyet to Kyoto, Blow for Campaign as Putin Jokes about
Global Warming,” The Mirror, September 30, 2003, p. 4.
2 Quoted in “Will Russia’s Heat Wave End Its Global-Warming
Doubts?” Time Magazine, August 2, 2010.
3 The present pH of the ocean is approximately 8, meaning that it is a
base. Acidification here means that the pH is decreasing, not that the
ocean will actually become acidic. For the ocean to actually become
acidic, its pH would have to drop below 7, which is very unlikely.
4 Numbers from www.climatecentral.org/news/noaa-2012-was-
warmest-and-second-most-extreme-year-on-record-15436. See also
Meehl et al. (2009), Relative increase of record high maximum
temperatures compared to record low minimum temperatures in the U.S,
Geophysical Research Letters, 36, L23701,
doi:10.1029/2009GL040736.
5 One frequently hears that increasing carbon dioxide will be good for
plants – i.e., “carbon dioxide is plant food.” Atmospheric carbon dioxide
is indeed a key ingredient in plant growth and, everything else being
equal, more carbon dioxide in the atmosphere would be expected to
increase the rate of plant growth. However, other changes, such as
changes in temperature and precipitation, are expected to offset the
benefits of increased carbon dioxide, particularly for warming of more
than a few degrees Celsius.
6 See Stanton and Ackerman (2007).
344

http://www.climatecentral.org/news/noaa-2012-was-warmest-and-second-most-extreme-year-on-record-15436

10
Exponential growth
◈
Before we continue our discussion of climate policy, we need to take a
detour to examine exponential growth, which may be the most important
term that you have never heard of. It touches many aspects of your life,
from the growth of credit card debt and housing prices to governing key
processes in biology, physics, economics, and, yes, climate change.
345

10.1 What is exponential growth?
First, a definition: exponential growth means that the rate of growth is
directly proportional to the present size. A good example of exponential
growth is the accumulation of money in a savings account. Imagine that
you deposit $100 into a bank account with an interest rate of 10
percent/year.1 After the first year, you receive interest equal to 10 percent
of the balance of $100, which is $10. This interest raises the balance of the
account to $110. After a second year, the interest is 10 percent of the
balance of $110, which is $11. This increases the balance to $121. Table
10.1 shows the growth of the bank account over 101 years.
Table 10.1 Calculation of the balance of a bank account with an initial
investment of $100 at an interest rate of 10 percent/year
Year Interest ($) Balance ($)
$100
1 10 110
2 11 121
3 12.10 133
4 13.30 146
5 14.60 161
6 16.10 177
7 17.70 195
346

(10.1)
8 19.50 214
⁞
100 125,278 1.38 million
101 137,806 1.52 million
This growth is exponential because the increase in the bank balance
during any year is proportional to the bank balance in that year. In fact,
anything growing at “x percent/year” is growing exponentially.
The key parameter in exponential growth is the rate of growth, or r.
For bank balances, credit cards, or mortgages, r is usually called the
interest rate, whereas in other contexts r may have other names (later in the
chapter, I will refer to r as the discount rate). Usually, r is expressed in
percent/year. Given a growth rate of r percent/year, an initial quantity P
will grow by a factor of 1 + r⁄100 in one year. So, after one year, the
quantity has grown to P(1 + r⁄100). For a bank balance of $100 and an
interest rate of 10 percent/year, the bank balance after one year is $100(1 +
10⁄100) = $100(1.1) = $110, the same answer we obtained earlier.
At the end of two years, the balance is P(1 + r⁄100)(1 + r⁄100). This is
simply the balance at the end of the first year, P(1 + r⁄100), multiplied by
another factor of 1 + r⁄100 to account for growth during the second year.
Thus, the bank balance at the end of the second year is $100(1 + 10⁄100)(1
+ 10⁄100) = $100(1.1)(1.1) = $110(1.1)2 = $121.
You may well be able to see a pattern here. After n years, an initial
investment of P will grow to a final value F:
This is the formula I used to generate the values in Table 10.1.
347

10.2 The rule of 72
When thinking about exponential growth, it is frequently useful to
consider the doubling time – the length of time that it takes for something
growing exponentially to double. From Table 10.1, for example, we see
that $100 invested at 10 percent/year will double in approximately seven
years to $200. After another seven years, the $200 doubles to $400, and
after another seven years, the $400 will grow to $800.
A simple way to estimate the doubling time is to use the rule of 72:
The doubling time is 72 divided by the growth rate (in percent/year).
Using this equation, we see that the doubling time at 10 percent/year is
72⁄10 = 7.2 years, a result that is consistent with Table 10.1. Note that
doubling time is a function only of growth rate. Thus, for a growth rate of
r, the doubling time is the same regardless of the size of the growing
quantity.
Using this rule, you can frequently do exponential growth problems
with pencil and paper – or even in your head. For example, let us put $100
in the bank at 7.2 percent interest in Year 2000. What is the balance in
Year 2100? The doubling time is 72⁄7.2 = 10 years, so that the balance
doubles every 10 years. Table 10.2 shows the balance at the end of every
decade.
Table 10.2 The balance of $100 invested in Year 2000 at an interest rate
of 7.2 percent/year
Year No. of doublings Value ($)
2000 – 100
348

(10.2)
2010 1 200
2020 2 400
2030 3 800
2040 4 1,600
2050 5 3,200
2060 6 6,400
2070 7 12,800
2080 8 25,600
2090 9 51,200
2100 10 102,400
After 100 years the $100 investment has grown to $102,400 –
illustrating the power of exponential growth. In equation form, an initial
investment of P has grown after n doublings to a final value F:
This is the equation I used to calculate the values in Table 10.2. And in a
pinch, I could have done the calculation with just pencil and paper.
We could also have used Equation 10.1 to calculate the balance in
2100: F = $100(1 + 7.2⁄100)100 = $104,587. This is very close to the value
calculated for 2100 by use of the doubling time, although the estimates
differ slightly. The difference results from the fact that the rule of 72 is
approximate, so calculations using it are almost always slightly off. In
most cases, though, the rule of 72 is accurate enough.
349

(10.3)
As another example, imagine investing $100 at an interest rate of 14.4
percent. How long would you have to leave this investment in the bank to
yield $1 trillion ($1012)? First, let us figure out how many doubling times
it would take. Using Equation 10.2, we can write the relevant equation as
Solving this equation2, we find that n = 33.2. In other words, an initial
investment of $100, doubled 33.2 times, yields $1 trillion. Now let us
calculate how many years that takes. At an interest rate of 14.4
percent/year, the doubling time is 72⁄14.4 = 5 years, so 33.2 doublings
takes 166 years. The growth of this investment is plotted in Figure 10.1.
Figure 10.1 Value of $100 invested at 14.4 percent interest as a function
of years invested.
Most of the accumulation occurs in the last few doubling periods. In
fact, $500 billion dollars, half of the total, is earned in just the last
doubling period, and 97 percent of the $1 trillion is earned in the last five
doubling periods (25 years). And, had the investment period been cut in
half (from 166 to 83 years), you would have $10 million at the end,
1⁄10,000th of what you earn for the entire time period. Thus, the
exponential growth is heavily weighted toward the very end of the
350

investment period. This has some important consequences, which we
investigate in the next section.
351

10.3 Limits to exponential growth
Where I live, College Station, Texas, the population growth rate peaked at
11 percent/year in the 1970s. At that rate, College Station’s population was
doubling every 72⁄11 = 6.5 years, or about 15 doublings per century. At
that growth rate, sustained over a century, College Station’s population
would have increased from about 37,000 in 1980 to 1.3 billion by 2080.
This is an impossibly high population and means that this growth rate
could not possibly be sustained for 100 years. In fact, considering the
practical limits of city growth, such high population growth is
unsustainable for much more than a decade or two. Thus, it should have
been obvious in the 1970s that the population growth rate of College
Station would decrease – and it has, to roughly 3 percent/year today.
A growth rate of 3 percent/year sounds much more sustainable, and it
is, but exponential growth is so stunningly fast that even seemingly low
growth rates can be unsustainable for more than a few centuries. For
example, at 3 percent/year, College Station’s population would approach 1
billion in about 300 years or so. This is also clearly unattainable in any
practical sense, and it says that, over the upcoming centuries, the
population growth rate of College Station must decline even further.
This brings us to the first important rule of exponential growth, best
expressed by economist Kenneth Boulding, “Anyone who believes
exponential growth can go on forever in a finite world is either a madman
or an economist.”3
A second rule of exponential growth is that, when the end comes, it
comes quickly. As an example, let us consider a resource consumed by
humans, such as water. Let us assume that in Year 2000, a community is
352

consuming 100 million gallons of water per year. As the community
grows, the rate of consumption doubles every ten years. Table 10.3 shows
this rapid increase in the rate of consumption over the twenty-first century.
Table 10.3 The amount of water consumed each year by a community,
assuming that consumption increases by 7.2 percent/year
Year Gallons per year
consumed
Fraction of supply
consumed (%)
2000 100 million 0.1
2010 200 million 0.2
2020 400 million 0.4
…
2070 12.5 billion 12.5
2080 25 billion 25
2090 50 billion 50
2100 100 billion 100
2110 200 billion 200
2120 400 billion 400
Note: The middle column shows the water consumed each year. Dividing
the rate of consumption by the maximum amount that can be supplied (100
billion gallons/yr) yields the fraction of the supply that is being consumed
(right-hand column).
The local reservoir that supplies the water can supply a maximum of
100 billion gallons of water per year. In Year 2000, the community is
353

using just 0.1 percent of the reservoir’s capacity. This is such a small
amount compared to the size of the supply that the community members
do not even consider the idea that they might one day run out of water.
As consumption increases over the twenty-first century, the
community uses a larger and larger fraction of the available water. This is
seen in Figure 10.2, which shows the time series of the fraction consumed.
By 2070, the community is consuming 12.5 percent of the supply, more
than a 100-fold increase since 2000 in the rate of consumption and in the
fraction of the supply used.
Figure 10.2 Consumption of water as a fraction (in percent) of total
supply, based on the data in Table 10.1.
That same year, a small group of activists claim that the world is on
the verge of exhausting its water supply. Most citizens dismiss this claim
out of hand: The community is using only 12.5 percent of the supply, and
it took 70 years of growth to go from 0.1 percent to 12.5 percent. Based on
common sense, the majority argue, it will be centuries before the water
supply’s limit is reached.
But the majority is wrong. As Figure 10.2 shows, around 2070
something remarkable happens: The consumption of water “turns the
corner and begins heading up rapidly.4 The reason is that, during
exponential growth, the increase each doubling period is equal to the
354

increase during all of the previous doubling periods, combined. Between
2070 and 2080, consumption will grow from 12.5 billion gallons per year
to 25 billion gallons per year. So the increase in consumption between
2070 and 2080 of 12.5 billion gallons per year is equal to the total increase
between 2000 and 2070. And it will only take two more doublings before
consumption is 100 billion gallons per year, or 100 percent of the supply.
Now let us assume that the community suddenly realizes that it is
going to reach its limits of consumption in a few decades. Through
Herculean efforts to develop and deploy new technology, the reservoir’s
capacity is increased by a factor of four – so it begins supplying 400
billion gallons of water per year.
That is an enormous increase – but it buys far less time than you
might think. If the growth rate remains constant, the community will be
consuming 100 billion gallons per year in 2100, 200 billion gallons per
year in 2110, and 400 billion gallons per year in 2120. This means that the
new water supply limit is reached in just two doublings or twenty years.
The lesson here is that, when things are growing exponentially, resource
limits may be closer than you think.
In reality, of course, economic factors come into play that will reduce
the growth rate. As limits to the resource are approached, the price should
go up, giving consumers a clear economic signal to change their behavior.
This will also encourage conservation, as well as technical development of
new sources and of substitutes.
Another good and recent example is the subprime housing crisis that
hit the United States in 2007. Beginning around 2003, house prices began
rapidly increasing, with price increases in many metropolitan areas of 20
to 30 percent/year. Some regions, such as Las Vegas, saw even more rapid
price increases (Figure 10.3). These growth rates correspond to doubling
times of a few years. At that rate, a $250,000 house would become a $2
355

million house in about a decade (3 doubling periods). Common sense tells
us that is impossible – at that rate, houses would rapidly become
unaffordable. Such growth in house prices had to stop, and soon.
Figure 10.3 Annual change in house prices in Nevada
(data obtained from the St. Louis Federal Reserve Web site; see
research.stlouisfed.org/fred2).
Although this conclusion may seem obvious in retrospect, at the time
people seemed to believe this exponential growth could go on forever.
Homebuyers were willing to pay ever-increasing prices and take out loans
they could not possibly afford because they believed that the value of their
house would always continue to appreciate. If they got in trouble, they
could always sell the house at a profit. Investors believed the same thing,
so they were willing to fund increasingly absurd mortgages.
Eventually, buyers with these insanely large mortgages began
defaulting on them, leading to a sudden, rapid collapse of the housing
market. Just a few years after houses were appreciating at 20 to 30
percent/year, their value was declining just as fast. Figure 10.3 shows how
356

http://research.stlouisfed.org/fred2

quickly this transition occurred. This is a frequent occurrence when things
are growing exponentially.
Probably the best-known warning of the dangers of exponential
growth came about 200 years ago from Thomas Malthus. Malthus was an
English cleric and scholar who argued that population grows exponentially
whereas food production grows in a linear fashion. Linear growth means
that the increase in each year is a fixed amount: Food production after n
years is F(n) = an + b, where a and b are constants and n is the number of
years. This is quite different from exponential growth, in which the
increase is a fixed fraction of the growing quantity. Mathematically, it is
easy to show that exponential growth will always eventually outpace linear
growth, and this led Malthus to conclude that an exponentially increasing
population would eventually outstrip the world’s ability to feed those
people, resulting in widespread starvation – what we now call a
Malthusian catastrophe.
Malthus was correct that population grows exponentially. However,
in the two centuries since Malthus’ prediction, technological developments
(e.g., development of fertilizers and pesticides) have allowed food
production to increase exponentially along with population. As a result, we
have not experienced this Malthusian catastrophe. We are not, however,
out of the woods yet. One of the main points of this chapter is that
exponential growth cannot continue indefinitely: Both food production and
population will eventually cease to grow exponentially. The question is
which one plateaus first. If exponential food production growth stops
before exponential population growth does, then Malthus’s prediction may
yet come true. If, however, population growth ceases before food
production growth, then Malthus may be forever wrong. Time will tell.
357

10.4 Discounting
In this section, I describe the financial concept of discounting, which plays
a key role in evaluating policy options for dealing with climate change.
358

(10.4)
10.4.1 The time value of money
Suppose you know today that you will incur an expense of $25,000 in
fifteen years. Such eventualities occur frequently in the world of corporate
finance and other business areas. How much would you be willing to pay
today to eliminate that future expense? One way to answer this is to
determine how much you would have to invest today in order to have
$25,000 in fifteen years. We can get the answer by rearranging Equation
10.1:
Here F is the expense, which will be incurred in n years, r is the interest
rate in percent, and P is the amount you need to invest today. Given an
interest rate of 5 percent, we need to invest about $12,000 today in order to
have $25,000 in fifteen years. In other words, we can view $12,000 today
as being equal to $25,000 in fifteen years.
One conclusion you can draw from this is as follows: Money in the
future is worth less to you than money today. That general conclusion is
probably obvious to most people, but this calculation allows us to answer
the question of how much less. The parameter r in these projects quantifies
the rate that money loses value as it recedes into the future – each year,
money loses r percent of its value. In these types of problems, r is
frequently referred to as the discount rate, and the value today of a future
expense or benefit is referred to as the present value. The process of
calculating the present value of a future cost or benefit is referred to as
discounting.
An example: What would you do?
359

You can use discounting to help make financial decisions. Imagine
you walk into an electronics store, searching for a new television.
You select one and are informed that you have two payment
options: You can get it “with no money down” and pay $1,100 in
one year, or you can pay $1,000 today. Which option do you
choose?
Note that $1,100 in one year has a present value of $1,100⁄(1
+ r⁄100), where r is the discount rate. If you choose a discount rate
of, say, 5 percent, then the present value is $1,047. This is more
than $1,000, meaning that $1,100 in one year is more expensive
than $1,000 today. You want to pay as little as possible for the
television, so you therefore choose to pay $1,000 today.
For a discount rate of 15 percent, the present value of $1,100
in one year is $956, so in that case $1,000 today is more expensive
than $1,100 in one year – and you would therefore prefer to pay
$1,100 in one year. If you choose a discount rate equal to 10
percent, then the present values are equal, and you would have no
preference about paying $1,000 today or $1,100 in one year.
In the policy debate over climate change, our choice is between
spending money today to reduce emissions of greenhouse gases, thereby
reducing the impacts of climate change in 50 to 100 years, or doing
nothing now and spending more money dealing with the impacts of
climate change in a few decades. So, for example, our choice might be to
spend $100 billion today or $1 trillion in 100 years.
Discounting allows us to quantitatively compare these two options. If
we assume a discount rate of 3 percent, then the present value of $1 trillion
in 100 years is 1012/(1.03100) = $52 billion. This is less than the alternative
360

of spending $100 billion today, so (from a purely financial perspective) we
would prefer to pay $1 trillion in 100 years than $100 billion today. This
type of analysis, in which you compare the present value of the costs and
benefits of various options, is a cost-benefit analysis, and economists
frequently use it to provide guidance to policymakers about alternative
policy options.
361

10.4.2 The discount rate
In both the television purchase and climate examples just given, the
answer we get is strongly dependent on the discount rate. In the climate
example, a discount rate of 3 percent yields the conclusion that we would
prefer to do nothing now and pay later to address the impacts of climate
change. However, if the discount rate were 2 percent, then the present
value of $1 trillion in 100 years is $138 billion, and we would rather pay
$100 billion today to reduce emissions.
So how do we determine the correct discount rate? The discount rate
really is a combination of two different judgments. First is what is known
as time discounting, which is the preference to consume now rather than
later. If offered $100 now or $100 in one week, just about everyone would
choose to get the money now. After all, why would you wait? Experiments
show that animals also exhibit this behavior. I know that if I give my dogs
the choice of having dinner now or in an hour, they tell me in no uncertain
terms that they want to eat now. In other words, most people (and dogs)
have a positive time discount rate: Goods and services now are worth more
than the same goods and services in the future.
The climate problem covers periods longer than a human lifetime. In
that case, the time discount does not represent our preference for us to
consume now rather than later. Rather, it represents our preference for us
to consume rather than future generations. Given that consumption can be
roughly equated to welfare, the time discount rate then expresses how
much we value our welfare above the welfare of future generations.
From a moral standpoint, most people agree that it is unethical to
place a higher value on our own welfare over that of future generations,
which implies that the time discount rate should be set to near zero.
362

Nonetheless, it is clear from our society’s actions, such as our low rate of
savings and our failure to address big problems facing future generations
(e.g., climate change, budget deficits), that we do indeed value our own
generation more highly, which implies a positive time discount rate.
The other part of the discount rate is known as growth discounting,
and it reflects the fact that a dollar means more to poor people than it does
to rich people. For example, if a billionaire is walking down the hall and
sees a $1 bill on the floor, would he stop to pick it up? Probably not – if
you have 1 billion dollars, another dollar does nothing to improve your
welfare. If you are living in poverty, however, you are most certainly
going to stop and pick up the $1 bill – it might be the difference between
having dinner that night or not.
In economics jargon, the utility of $1 to the billionaire is much lower
than the utility of $1 to the person living in poverty. In the case of climate
change, we expect future generations to be richer than we are, just like we
are richer than those living 100 years ago. And because future generations
are richer, they will be better able to pay costs associated with climate
change than we are. This suggests a preference for our generation to push
the costs of addressing climate change onto richer, future generations. This
preference is expressed in the growth discount rate, which is the rate at
which the utility of money – how much each dollar means to society –
declines with time as the world gets richer.
The discount rate used in present-value calculations is determined by
combining the time and growth discount rates. Unfortunately, the choice of
both time and growth discount rates is as much of a value judgment as an
objective fact, and there are wide disagreements about what discount rate
to use. Some economists argue the discount rate should be near zero,
whereas others argue for higher values such as 4 percent.
363

This makes a huge difference in climate problems. For future impacts
of $1 trillion in 100 years, the different discount rates yield present values
of $1 trillion (for a 0 percent discount rate) or $19 billion (for a 4 percent
discount rate). The policy implications of these hugely different estimates
of the present value of climate impacts are completely different.
364

10.5 Putting it together: The social cost
of carbon
Imagine that you emit a ton of carbon dioxide to the atmosphere. As
discussed in the first half of the book, this ton of carbon dioxide will warm
the climate over many thousands of years – until the carbon cycle
completely removes it from the atmosphere. As described in Chapter 9,
this warming imposes costs on our society. For example, increases in sea
level require construction of expensive defenses against the sea (e.g., sea
walls) or relocation of communities being inundated. And this is just one
cost from carbon dioxide; there are also costs associated with rising
temperatures, changing precipitation patterns, ocean acidification, etc.
Calculating the total cost imposed on society by this ton of carbon dioxide
requires adding up all of the costs from all of the impacts.
Once you know the total cost of this ton each year, the costs can be
discounted back to today. As an example, imagine that the ton causes $1 of
damage every year for 300 years. Assuming a discount rate of 3 percent,
the present value of the cost of impacts during the first year is $1⁄1.03, the
second-year’s damages is $1⁄1.032, and year n’s damages is $1⁄1.03n.
Figure 10.4 shows the present value of $1 of climate damages for two
discount rates, as a function of when the damages occur. The figure shows
that, as the damages recede into the future, the present value of the
damages declines rapidly. For the higher discount rate, costs beyond about
2200 (200 years from now) have a present value of zero. For the lower
discount rate, the present value drops off more slowly and a small fraction
of the damages in 2300 are contributing to the present value.
365

Figure 10.4 The present value of $1 of climate damages, as a function
of the year the damages took place. The calculation is shown for two
discount rates.
We can calculate the total cost to us today from the emission of this
ton of carbon dioxide by summing the discounted costs for every year.
This total cost is frequently referred to as the social cost of carbon, and in
this simple example it equals $33 and $94 for the discount rates of 3
percent and 1 percent, respectively.
Sophisticated expert estimates of this quantity from a recent
assessment are listed in Table 10.4. There are several things worth noting
in this table. First, seemingly small changes in the discount rate lead to
large changes in the social cost of carbon. This emphasizes the central role
of the discount rate in the climate-change policy debate.
Table 10.4 Estimates of the social cost of carbon (in 2007 dollars) for
different discount rates. The 5%, 3%, and 2.5% columns are the averages
of the estimates from a large number of economic models; the 3%, 95th
percentile column is the value exceeded by 5% of the estimates of the
estimates using a discount rate of 3%.
366

Discount rate
Year 5% 3% 2.5% 3%, 95th
percentile
2010 $40 $121 $191 $330
2015 44 139 213 400
2020 44 158 239 473
2025 51 176 257 528
2030 59 191 279 584
2040 77 228 319 705
2050 99 261 360 811
Note: This table is the value per ton of carbon; the value per ton of carbon
of carbon dioxide is 1/3.67 of this value.
Source: Technical Support Document: Technical Update of the Social
Cost of Carbon for Regulatory Impact Analysis, Under Executive
Order 12866
(www.whitehouse.gov/sites/default/files/omb/inforeg/social_cost_of_carbon_for_ria_2013_update
This is an update of a 2010 analysis
(www.epa.gov/otaq/climate/regulations/scc-tsd ).
Second, the social cost of carbon rises throughout the twenty-first
century. The reason for this is that damage from climate change is non-
linear: each degree of warming produces more damage than the previous
degree. Thus, as we add carbon dioxide to the atmosphere, the cost of the
damage from each additional ton is more than the cost from the previous
ton.
367

http://www.whitehouse.gov/sites/default/files/omb/inforeg/social_cost_of_carbon_for_ria_2013_update

http://www.epa.gov/otaq/climate/regulations/scc-tsd

Third, comparing the 3% column and the 3%, 95th percentile column
gives some idea of how wide the range of estimates are for a single
discount rate. The 3% column is the average of all of the estimates, while
the 3%, 95th percentile column is the value exceeded by 5 percent of the
estimates using a 3 percent discount rate. The large difference between
these values says that, even at a single discount rate, there is a wide range
of estimates of the social cost of carbon, with some models estimating
much larger values than others.
The wide range of estimates in Table 10.4 underscores two primary
difficulties in estimating this quantity. First, the value you get is sensitive
to the choice of discount rate, and there is little agreement among
economists as to what the right value is. Second, putting a dollar price on
the impacts of climate change requires a set of linked predictions, all of
which are highly uncertain. We need to predict how the climate will
change in individual regions, how that climate change will affect people
and ecosystems, and how people and ecosystems will in turn respond to
these climate changes. Then, a dollar value must be assigned to the
impacts.
For some impacts, a dollar value is easy to assign. For goods and
services sold on the open market (e.g., agricultural products, farmland, and
coastal property), the market value can be used as an estimate of their
social value. Other impacts are much more difficult to value. A good
example is the existence of polar bears. They have little market value, but
(to me, at least) they do have social value – I believe that driving them to
extinction would be a great loss to humanity. Most people would agree
that there is some value in their continued existence, but there might be
great disagreement over exactly how much.
The difficulty in calculating the overall costs of climate impacts is
responsible for the large spread of estimates in Table 10.4. However,
368

despite the large range of estimates, there are a few things we can conclude
with confidence:
1) The social cost of carbon is not zero. Every credible economic
analysis of climate change found has found that there are costs from
the emissions of greenhouse gases and the associated climate change.
2) While there is uncertainty in our estimates of the social cost of
carbon, very large costs cannot be ruled out. Such large costs are
associated with dire climate change impacts, and the possibility of
this bleak future drives much of the concern about the climate
problem.
Despite the uncertainty in this quantity, regulators must nonetheless assign
a single value for the social cost of carbon in order to use it in cost-benefit
analyses and other regulatory decisions. In 2013, the Obama
Administration selected $132 per ton of carbon ($36 per ton of carbon
dioxide) as our best estimate of today’s social cost of carbon.
369

10.6 Chapter summary
Exponential growth means that the rate of growth is directly
proportional to the present size. Anything growing at “x
percent/year” is growing exponentially.
A quantity P growing at r percent/year will grow to P(1 + r⁄ 100)n
after n years. The time for a quantity growing exponentially to
double is frequently referred to as the doubling time; it is
approximately equal to 72⁄r, and this shortcut is known as the “rule
of 72.”
Exponential growth tends to be end loaded, meaning most of the
growth occurs at the end: 50 percent of the growth occurs during
the last doubling period, and 97 percent of the growth occurs
during the last five doubling periods.
Exponential growth cannot go on forever, and when it ends, it often
ends abruptly.
Discounting refers to the process of calculating the present value of
some future expense or benefit. Such calculations require a
discount rate, which is the rate at which money loses value in the
future. In general, money in the future is worth less than money
today.
Cost-benefit analyses compare the present values of the costs and
benefits of a range of policy options. The best option (from a
financial point of view) is the one with the largest present-value net
benefit.
370

The discount rate is determined by two judgments: the time
discount rate, which is our preference for consuming now rather
than later, and the growth discount rate, which reflects the fact that
future generations are expected to be richer so they can pay a
bigger share of the costs. There are vigorous disagreements over
what discount rate we should use.
The social cost of carbon is the cost, discounted to today, of the
future impacts due to the emission of one ton of carbon dioxide to
the atmosphere.
The social cost of carbon is a highly uncertain quantity, due to
uncertainty in the discount rate and from difficulty in assessing the
monetary value of climate change impacts. Despite the uncertainty,
we can be confident that there are net costs to climate change, and
these costs could potentially be very large. In 2013, the Obama
Administration selected a value of $132 per ton of carbon as our
best estimate of the social cost of carbon.
371

Additional reading
Much of the material in this chapter was motivated by several YouTube
videos of lectures by Dr. Albert Bartlett on exponential growth. I highly
recommend watching them. Dr. Bartlett also wrote numerous papers on
exponential growth, many of which are collected in his book, The
Essential Exponential! See www.andrewdessler.com/chapter10 for links to
Dr. Bartlett’s material as well as other related resources.
372

http://www.andrewdessler.com/chapter10

Terms
Cost-benefit analysis
Discounting
Discount rate
Doubling time
Exponential growth
Growth discounting
Interest rate
Malthus
Present value
Rule of 72
Social cost of carbon
Time discounting
373

Problems
1. You invest $1 at a 10 percent interest rate for fifty years.
a) Use Equation 10.1 to calculate how much you have after fifty
years.
b) How many doubling periods does the investment experience?
c) Use Equation 10.2 to calculate how much you have after fifty
years.
2. You invest $50 at a 7 percent interest rate for thirty years.
a) Use Equation 10.1 to calculate how much you have after thirty
years.
b) How many doubling periods does the investment experience?
c) Use Equation 10.2 to calculate how much you have after thirty
years.
3.
a) How many doubling periods do you have to wait for 1 cent to
grow to $100 trillion? (Calculate to the nearest integer.)
b) At an interest rate of 7 percent, about how long does it take for
that many doublings to occur?
4. Would you rather pay $1 trillion dollars of damages from and
adaptation to climate damage in fifty years or pay $50 billion dollars
today to reduce emissions and avoid the climate change? Use
discount rates of 0 percent, 2 percent, 4 percent, 6 percent, and 8
percent.
374

5. You go into a big-box electronics store to buy a flat-screen
television. You have two options: pay $1,400 today or $1,450 in one
year. Which do you choose? You have to estimate a discount rate to
do this. How did you choose your discount rate?
6. Lotteries often give you the option of taking a lump-sum payment
now or a fixed amount every year for, say, twenty-five years. For this
question, assume that the lump-sum payment is $3 million and the
yearly payments are $250,000 each year for twenty-five years. The
first payment is made immediately, so it is not discounted, and
subsequent payments are made every year thereafter.
a) Find the discount rate where the present value of the twenty-
five-year cash stream is equal to the lump-sum payment.
b) If your discount rate is higher than this, should you take the
lump-sum payment or the period payments?
7. In the National Football League draft, a pick in this year’s draft is
worth a pick in a lower round in a future draft (e.g., you might trade a
second-round pick in this year’s draft for a first-round pick in next
year’s draft). Explain how this is consistent with the concept of
discounting.
8.
a) Imagine you have a dollar bill. If you double it, you have two
bills. If you double again, you have four bills. If you double again,
you have eight bills, and so on. Given that a bill is 0.1 mm thick,
how many doublings do you have to go through before you have a
stack that reaches from the Earth to the moon? (The moon is
360,000 km away.)
375

b) How many doublings do you need to get a stack that goes
halfway to the moon?
c) How many doublings to you need to get 1 percent of the way to
the moon?
9.
a) Consider the choice between paying $10 million today to reduce
emissions that cause climate change or $1 billion in 100 years to
adapt to a changing climate. What would the discount rate have to
be in order for these two choices to be equal?
b) Using that same discount rate, what would be your preference if
the expense was in 50 years instead of 100?
10. You are inside the Houston Astrodome, in the rafters just below
the roof, 160 ft above the field. A wizard puts a tiny magic drop of
water on the pitcher’s mound, and the drop starts doubling in volume
every minute. After 100 minutes, there is 5 ft of water on the field,
and the depth continues to double every minute. How many minutes
do you have before the water reaches you?
11. Calculating the cost of climate change.
a) If the economy grows at 3 percent/year, how many times larger
than today will it be after 100 years?
b) Imagine that addressing climate change reduces economic
growth from 3 percent to 2.9 percent over the century. How much
smaller is our GDP in 100 years?
c) How many additional years of growth at 2.9 percent need to
occur until the GDP is as large as 100 years of growth at 3 percent?
376

d) Put yourself in the shoes of a future citizen: Given how much
richer people will be in 100 years (that is the answer to part a),
should we be concerned about the loss of wealth due to a reduction
in growth from 3 percent to 2.9 percent that we calculated in part
b?
12. Imagine that you can save the polar bears if you pay a fee every
year. If you do not pay the fee, they go extinct. How much would you
pay each year to keep polar bears alive?
1 We are assuming here that the interest is compounded annually. Most
bank accounts and credit cards calculate interest monthly, meaning that
the balance is increased each month by the balance times the annual
interest rate divided by 12. Monthly compounding grows the balance
faster for a given interest rate than annual compounding.
2 There are several ways to solve this equation. One way is to iterate.
This means that you guess a value for n and plug it into your calculator.
If that n produces too large a number, then reduce your estimate of n
and repeat the process; if it produces too small a number, then do the
opposite. With a bit of practice, you can quickly converge on the correct
answer with five to ten guesses. You can also solve the equation
algebraically by taking the log of both sides of the equation and
rearranging to solve for n.
3 Quoted in Deffeyes (2006).
4 If the exponential were infinite, the choice of where the exponential
turns the corner is arbitrary and can be selected by carefully choosing
the scale. However, when it is exponentially growing consumption of a
fixed resource, the region where the curve turns the corner is real and set
by the maximum value of the resource.
377

11
Fundamentals of climate
change policy
◈
In the previous chapters of this book, we have seen that 1) the Earth is
warming, 2) most of the recent warming is very likely due to human
activities, 3) warming over the next century will likely be a few degrees
Celsius, and 4) such warming carries with it a risk of serious, perhaps even
catastrophic impacts for humans and the planet’s ecosystems.
Given those facts, what shall we do about climate change? Science, it
turns out, is just one of several factors needed to answer this question.
Deciding what to do also requires information about the options available
to us to respond to climate change, and the costs, benefits, and risks of
each option. In addition, we must consider not just monetary costs but also
the moral implications of each policy. In this chapter, I will outline the
various options available to us to address climate change.
Our responses to climate change can be broadly split into three
categories: adaptation, mitigation, and geoengineering. Adaptation means
responding to the negative impacts of climate change. If climate change
causes sea-level rise, an adaptive response to this impact would be to build
seawalls or relocate communities away from the encroaching sea.
378

Mitigation refers to policies that avoid climate change in the first place,
thereby preventing impacts such as sea-level rise from occurring. This is
accomplished by reducing emissions of greenhouse gases, usually through
policies that encourage the transition from fossil fuels to energy sources
that do not emit greenhouse gases.
Geoengineering refers to active manipulation of the climate system.
Under this approach, our society would continue adding greenhouse gases
to the atmosphere, but we would intentionally change some other aspect of
the climate system in order to cancel the warming effects of the
greenhouse gases. For example, we could engineer an increase in the
albedo of the Earth. If done correctly, this could stabilize the global-
average climate despite continuing emissions of greenhouse gases. In the
rest of this chapter, we explore each of these options in detail.
379

11.1 Adaptation
As temperature, precipitation, sea level, and other components of our
climate change, we can adapt our way of life to adjust to these changes.
There are several advantages to relying on adaptation as the main response
to climate change. First, because many of the worst impacts of climate
change will occur in the second half of the twenty-first century, adaptation
allows us to wait for decades before we must start adapting. There are
clear advantages to doing this – for example, it allows us to resolve
uncertainty on how climate will change and focus our efforts on the severe
impacts, without wasting any effort dealing with impacts that do not
materialize.
Second, if the past is any guide, we can expect future generations to
be richer than we are and better able to bear the costs of adaptation (this is
the factor that growth discounting, discussed in the last chapter, accounts
for). Third, many of the adaptations necessary to address climate change
will simultaneously benefit society in other ways. Decreasing a
community’s vulnerability to sea-level rise caused by climate change will
also decrease the vulnerability to extreme sea-level events caused by
hurricanes and other severe storms. Conserving water to address decreased
freshwater availability will also decrease the community’s vulnerability to
droughts or to increased demand caused by population growth. An
improved public health infrastructure designed to head off disease
outbreaks in a warmer world would also help decrease a society’s
vulnerability to pandemic flu and other nonclimate public health issues.
Finally, of all of the possible responses to climate change, adaptation
requires the least intrusion of government into the private lives of
380

individual citizens. In fact, strictly speaking, adaptation requires no
government intervention at all. If sea level rises and the government does
nothing, individual citizens will not just sit there and be submerged, they
will make individual decisions about how to respond: whether to build
coastal defenses, move to higher ground, etc. If climate changes in
agricultural areas, affected farmers can decide whether to relocate, change
farming practices, etc. Because of this, those philosophically opposed to
government regulation find adaptation to be a less serious infringement on
personal liberty than mitigation or geoengineering.
However, some of these advantages are largely illusory. Waiting until
the impacts of climate change are obvious is much more expensive than
adapting in advance. For example, if you are building an airport on the
coast that may last 100 years, it is better to spend money today to build
sea-level rise into the design than it is to wait for the airport to flood in 70
years and then face much higher costs of fixing the damage from the flood
as well as making the infrastructure less vulnerable to future floods (e.g.,
sea walls, levees).
And the idea that adaptation can be a wholly local response is also
largely illusory because adaptation takes huge resources. For example,
building a sea wall to protect a community against rising sea levels is
expensive, and most individuals and even most individual communities,
simply cannot afford to build one. Because of this, it is generally agreed
that the significant adaptation efforts require national governments or
international institutions to provide resources. And such national or
international assistance to a local community often makes sense. If sea-
level rise submerges Miami, and the resulting economic disruption hurts
the entire U.S. economy, then the U.S. federal government might be
justified in paying for seawalls to prevent that from happening. Or if
climate impacts in China threaten to destabilize the world economy,
381

international assistance to help the Chinese deal with the impacts may be
appropriate.
In addition to direct aid, governments can also implement regulations
to encourage citizens to adapt to a changing climate. Regulations
promoting water conservation, for example, would help communities adapt
to decreased freshwater availability. Governments can also eliminate
existing regulations that encourage us to be poorly adapted to the present
climate and that increase our vulnerability to climate change. A good
example is flood insurance. People love to build houses near bodies of
water, such as the ocean. However, the downside of this is that flooding
may occasionally destroy their houses. Without flood insurance, many
people would find it too risky to build in flood-prone areas because they
could not afford to have their houses destroyed. With flood insurance,
however, people can afford to live in flood-prone areas; if their houses are
destroyed by flood, the insurance covers the loss. In this way, flood
insurance actually encourages people to build where it is going to flood.1
A third way government policy can facilitate adaptation is by
providing reliable information about climate change, as well as possible
responses, including technical assistance. Such information would help
people plan for climate change and adapt in advance rather than waiting
for disaster to strike and then dealing with climate change.
There are also significant disadvantages to relying on adaptation as
the main response to climate change. Because adaptation to climate change
requires resources, the effects of environmental disruption are felt most
strongly by the poorest and most vulnerable in any society. This was ably
demonstrated when Hurricane Katrina hit New Orleans in 2005. When the
storm hit, the wealthier residents of New Orleans, those with resources
such as credit cards and automobiles, simply left town. The poorest
residents of the city, who lacked resources to evacuate, were stranded in
382

New Orleans and made up the vast majority of those killed. The poor also
tended to live in more poorly constructed housing, frequently located in
marginal areas, such as swampy or low-lying, flood-prone land.
Once the storm passed, the wealthier residents of New Orleans had
the resources to reconstruct their lives – either to return and rebuild or to
start anew somewhere else. The poorer residents had to rely on
government assistance to rebuild their lives. Because the United States is
rich, it was able to provide assistance such as money and temporary
housing that kept Katrina from being the much larger humanitarian
disaster it would have been for a poorer country.
We also see connections between vulnerability and wealth in two
recent earthquakes. In early 2010, a magnitude-7.0 earthquake ravaged
Haiti – it killed more than 200,000 people, made 1 million people
homeless, and heavily damaged much of the country’s built structures. A
few weeks later, a magnitude-8.8 earthquake hit Chile. Although this
earthquake was about 500 times stronger, it killed just a few hundred
people. Much of the difference in death toll can be attributed to Chile’s
greater wealth. Richer countries have the luxury of spending more money
on infrastructure, so buildings in Chile were built to more stringent
standards and most did not collapse during the quake.
The connection between the severity of climate change impacts and
wealth means that adaptation fails a fundamental fairness test. The world’s
rich economies have built their wealth by consuming massive amounts of
energy, which means that these economies are responsible for most of the
global warming over the past two centuries. Yet their very richness allows
these countries to deal most effectively with the impacts. The poorest
countries in the world are responsible for very little of the greenhouse
gases in our atmosphere today, yet they are least capable of dealing with
the impacts.
383

Because of this, an adaptation-only response is frequently viewed as
morally problematic because it abandons the poorest people in the world to
the impacts of climate change that they did not cause. To the extent that
adaptation is necessary, there is general agreement in the international
community that rich countries must help the poorer countries of the world
adapt to the impacts of climate change.
Although the rich of the world have many more resources than the
poor to deploy against the impacts of climate change, do they have
enough? That depends on your definition of “successful” adaptation. If
you define successful adaptation as minimal survival of the human species,
then humans can successfully adapt to almost any climate change. After
all, we are a resilient species and it is hard to believe that any climate
change will lead to our extinction.
A more realistic standard of success is maintenance of our standard of
living. The challenge with this metric is that there is no single agreed-upon
way to measure it. At its most restrictive, the standard of living can be
equated to the amount of goods and services consumed (e.g., GDP per
person). A more expansive definition might also include the value of
nonmarket activities such as the opportunity to hike in the same pristine
wilderness you did as a child, the satisfaction of seeing the leaves in New
England change colors every fall, or appreciation for the existence of polar
bears, even if you never actually see one. None of these are fully
accounted for in GDP.
Whereas GDP is relatively easy to measure, estimating the value of
nonmarket activities is much more difficult. Most of us agree that a world
with polar bears is more valuable than a world without, but exactly how
much more? Because there is no universal answer to this question, a
particular strategy for adaptation might be viewed as a success by someone
who puts less value on polar bears and a failure by someone who puts
384

more. This can lead to starkly different views about our ability to adapt to
climate change.
Despite this, there are a few general rules about successful adaptation
that we can have high confidence in. First, the more warming we have, the
harder it will be to successfully adapt – regardless of our definition.
Second, for a given amount of climate change, how we define success in
adaptation plays a key role in determining our ability to successfully adapt.
The more expansive the definition, that is, the more value we place on
maintaining the environment in its present state, the harder it will be to
successfully adapt. In fact, for those who put the highest value on
maintaining the present environment, climate change already experienced
combined with committed warming over the next few decades may be so
large that it is no longer possible for us to successfully adapt.
The bottom line on adaptation: Because of lags in the climate
system, as well as lags in the economy, some future climate change is
unavoidable. To the extent that this warming cannot be stopped, we must
adapt to it. Thus, adaptation must be a part of our response. However,
relying entirely on adaptation as our response is problematic. Adaptation
requires resources, and many of the world’s poorest inhabitants have few
resources and therefore little ability to adapt. As a result, even low-to-
moderate climate change will impose harsh impacts on these people. Many
view this as morally unacceptable because the world’s poor have
contributed little to the problem of climate change.
Rich countries have more resources to address the problem, and for
low-to-moderate climate change they may well be able to successfully
adapt. But if climate change is at the upper end of the range of predictions
discussed in Chapter 8, even rich countries may not have enough resources
to adapt. Thus, relying entirely on adaptation as our only response to
climate change is a titanic gamble for the rich that climate change over the
385

next century will not be severe. For the poor, it is no gamble at all –
climate change will impose significant hardships. All of this, of course,
depends to some extent on the definition of successful adaptation.
As a result, adaptation-only policies are not seriously considered in
the climate policy debate, and there is wide agreement that mitigation must
be part of our solution to the problem of climate change.
386

11.2 Mitigation
Mitigation refers to reductions in emissions of carbon dioxide and other
greenhouse gases, thereby avoiding the impacts of climate change by
preventing the climate from changing in the first place. Because relying
entirely on adaptation is a risky strategy, most policymakers view
mitigation as the centerpiece of any long-term climate change policy.
There are several approaches that could be used to reduce emissions, and I
discuss the range of options available in this section.
Before we get into the details of mitigation, it is important to make
clear the size of the required reductions. As we will discuss in Chapters 13
and 14, almost every country in the world has agreed with the judgment
that warming of more than 2°C above preindustrial temperatures would be
considered dangerous. Stabilizing at or below this temperature would
require a reduction in the global emissions by the middle of the twenty-
first century of greenhouse gases by 50 to 80 percent below today’s
emissions levels and reaching near-zero emissions later in the century.
This is a challenging but achievable target, and we will discuss it in more
detail in Chapter 14.
In Chapter 8, we explored the factors that control emissions of
greenhouse gases: population, affluence, and technology. Thus, we can
recast the problem of reducing emissions into the problem of reducing one
or more of these factors until emissions reach a desired value. The first
factor is the world’s population. With fewer people on the planet
consuming goods and services, emissions would certainly decrease. Some
societies have already implemented policies to actively influence the size
of their population. China, for example, adopted a “one-child policy,”
387

which limits the number of children a family can have to one, although
there are many exemptions. This policy has significantly reduced China’s
population growth rate, although the total population is still increasing.
Reducing emissions through population control would require more
than just a reduction in the rate of population growth – it would require a
significant reduction in the actual number of people on the planet. Such an
effort would conflict with deeply held religious, social, and cultural
traditions surrounding reproduction and family size in many countries. It
also creates other demographic and societal problems – as China has
discovered in response to its one-child policy. As a result, efforts to
combat climate change by using policies explicitly targeted at reducing the
Earth’s population are viewed as politically unachievable, and there are no
serious discussions of this approach.
A second option is to reduce the world’s consumption of goods and
services. If each person consumed less, the amount of energy consumed,
and therefore emissions, would decrease. Like population, solving the
climate problem through consumption would require not just stopping
growth of consumption but deep reductions in it. There are several
problems with solving climate change this way. First is a political problem
– people equate consumption with well being. That is why all politicians
strive for increased consumption (which they call economic growth), and
no politician who wants to keep his job would agree to a policy that
steeply reduces it. Thus, reducing consumption is something that most
countries simply will not agree to.
Then there is the two billion or so of the world’s poorest inhabitants
who live in extreme poverty, with incomes of a few dollars per day. These
people do not have the basic necessities of life – food, clean water, shelter
– and lifting them out of poverty requires economic growth, which
translates into increasing consumption. Efforts to limit consumption to
388

address climate change might mean preventing these people from escaping
poverty. Such an outcome is generally viewed as morally unacceptable:
We cannot solve climate change on the backs of the world’s poorest
people.
Thus, like population control, there are no serious efforts to address
climate change by reducing the world’s level of consumption. The rich
world does not want to do it, and it is ethically problematic to impose such
a policy on the world’s poor.
If neither population nor consumption has any chance of being
reduced, then by process of elimination it is the technology term, also
referred to as the greenhouse-gas intensity, which must be reduced in order
to reduce emissions. As I discussed in Chapter 8, the technology term is a
measure of how much greenhouse gas is emitted per dollar of GDP, which
itself can be broken into two constituent terms: the energy intensity, a
measure of how much energy it takes to generate $1 of GDP (J/$), and the
carbon intensity, a measure of how much greenhouse gas is emitted to
generate a Joule of energy (CO2/J). If your memory of this is hazy, you
may want to review Chapter 8.
Reducing greenhouse-gas intensity therefore requires reducing energy
intensity, carbon intensity, or both. Energy intensity is determined to a
large extent by the efficiency with which the economy uses energy.
Today’s society wastes a tremendous amount of energy, and improving our
energy efficiency would not only reduce our emissions of carbon dioxide
but would have many co-benefits, such as saving us money and reducing
air pollution. Because of the co-benefits, energy-efficiency improvements
would make sense even if climate change were not a problem.
Can efficiency improvements lead to large enough reductions? To
reduce emissions by 50 to 80 percent over the next few decades, which is
about what is required to stabilize the climate with less than 2°C of
389

warming, would require reducing emissions by approximately 2
percent/year. If the world’s total GDP (the product of population and
affluence) grows by 3 percent/year, then energy intensity would need to
decline by 5 percent/year or so to achieve the necessary reductions in
emissions.
Historically, energy intensity has decreased at roughly 1 percent/year.
This can likely be maintained, and some improvement may be possible,
but most experts do not believe that rates of decline of energy intensity of
5 percent/year for several decades are realistic. So although improvements
in energy efficiency can contribute to emissions reductions and are
something we should be doing now, they are likely only going to play a
supporting role in solving the climate problem. We therefore conclude that
it is reductions in the carbon intensity term, the amount of carbon dioxide
emitted per Joule of energy generated, that are required to stabilize the
climate.
390

11.2.1 Technologies to reduce carbon intensity
Reducing carbon intensity is code for switching from conventional
combustion of fossil fuels to energy sources that do not release greenhouse
gases – often referred to as carbon-free or climate-safe energy sources.
These include nuclear energy, carbon capture and sequestration, and
energy sources known as renewable energy, because these energy sources
are not depleted when utilized: primarily hydroelectric, solar, wind, and
biomass energy.
Solar energy is one of the most frequently discussed renewable
energy sources. There are actually two different ways to generate energy
from sunlight: solar photovoltaic or solar thermal methods. Photovoltaic
energy is the most common form of solar energy, and you can see it in
operation in the form of solar panels located on houses or buildings (and
also on satellites and the Space Station). It takes advantage of the fact that,
when exposed to light, certain materials such as silicon produce electricity.
Solar thermal energy, in contrast, uses mirrors to concentrate sunlight on a
working fluid (such as an oil, molten salt, or pressurized steam), heating it
to several hundred degrees Celsius. This hot fluid boils water and
generates high-pressure steam that is used to turn a generator, producing
electricity.
Solar energy is in many respects the Holy Grail of renewable energy.
As we calculated in Chapter 4, the amount of solar energy falling on the
planet is staggering – more than 100,000 TW. This is an enormous amount
of energy compared to the amount humans consume, about 15 TW. There
are, however, problems with the large-scale adoption of solar energy. One
is intermittency – the Sun shines only during daytime and when not
obscured by clouds. Thus, solar energy may require additional
391

mechanisms to ensure the reliable twenty-four-hour availability of power
consumers expect. Another is the area required to generate solar energy.
Taking into account the intermittency and other efficiency issues, solar
energy can supply power at a level of approximately 10–20 W/m2. To
satisfy all human energy needs would therefore require roughly 1 million
km2 to be covered with solar energy collectors, corresponding to 0.2
percent of the Earth’s surface. Although this is a large area, it is
comparable to the total area covered by cities, so there is no reason to
believe that it is impossible for humans to construct the number of
collectors needed for that much solar energy.
Another frequently mentioned renewable energy source is wind. This
is a mature technology – the Dutch have been using wind energy for
hundreds of years to do useful work, such as pumping water. Today’s
electricity-generating windmills, often referred to as wind turbines, are
quite a bit larger and more sophisticated. The largest ones are 130 m tall,
the same as a forty-story building, with 125-m blades. A single one of
these wind turbines can generate as much as 6 MW of power, so that a few
hundred can replace a conventional fossil-fueled power plant.
Wind also has the problem of intermittency. The wind does not blow
everywhere nor does it blow all the time, so, on average, the 6-MW wind
turbine actually produces roughly one-third of its maximum value. Taking
into account the intermittency of wind, as well as the fact that windmills
must be spaced apart so that they do not interfere with each other, we find
that a wind farm generates power at a level of approximately 2 W/m2.
Therefore, to satisfy human energy requirements would require covering
approximately 1.5 percent of the Earth’s surface area with wind farms
containing a few million windmills. It should be noted that putting up
windmills does not preclude using the land simultaneously for other
activities, such as agriculture.
392

Although it would undoubtedly be an enormous undertaking to
construct the number of solar collectors and windmills needed to produce
that much solar or wind power, human industry has produced amazing
feats in the past. During World War II, for example, the countries of the
world produced hundreds of thousands of airplanes, tens of thousands of
tanks, thousands of ships, and massive amounts of other equipment of war.
Compared to that, putting up a few million windmills does not sound so
hard.
Wind and solar energy have been growing rapidly over the recent past
and are emerging as important contributors to our energy supply.
However, these sources remain (generally) more expensive than electricity
from fossil fuels, and the intermittency problem is still being worked out.
Thus, we are not yet on the verge of a wholesale transition of our energy
supply to these renewable sources.
Biomass energy is another renewable option; it refers to the process
of growing crops and then burning them to yield energy. Because the
carbon dioxide released from burning biomass was absorbed from the
atmosphere during the growth of the plant, there is no net increase in
carbon dioxide in the atmosphere. It is an intuitively attractive energy
source, but there are several issues that must be considered. First, the rate
of photosynthesis limits the power generated by biomass to roughly 0.6
W/m2 of farmed land. Thus, to generate 15 TW would require that 15
percent or so of the land surface be devoted to growing biomass for energy
– comparable to the area presently under cultivation today.
The enormous land requirement is problematic. We know from
experience that much of the additional land will come from clearing forest.
This deforestation releases carbon dioxide into the atmosphere, and it
causes a host of other local environmental impacts, such as loss of native
biodiversity and ecosystem degradation. The second problem is that the
393

farming methods used to grow the biomass have to be carefully
considered. Production of fertilizer, for example, requires large inputs of
energy, mainly from fossil fuels. If fertilizer is used in the growth of the
biomass, it might take as much fossil fuel energy to grow the biomass as is
saved by burning the biomass to produce energy.
Finally, it is becoming clear that using food, such as corn, as
feedstock for biomass energy severely stresses the food supply. The
increased competition for food raises food prices, an impact
disproportionately felt by the poor. The hope is that a technological
breakthrough will allow us to produce energy from waste biomass that
does not have other uses, such as the waste from corn processing (e.g.,
corn stalks, corn cobs) or cellulosic biomass such as switch grass. Many
scientists are currently working on methods to produce biomass energy
from these waste sources. Despite these difficulties, biomass, particularly
in the form of corn-based ethanol, already provides a few percent of U.S.
motor fuels and is slated by act of Congress to become a major source of
automotive fuel in the United States over the next decade.
Biomass energy systems are a promising technology, but any biomass
system must be carefully constructed from end to end to ensure that carbon
emissions are actually reduced (e.g., limiting how much fertilizer is used,
where the land comes from). In addition, new technologies that allow
biomass energy to be extracted from nonfood biomass must also be
developed. Thus, large-scale biomass energy production may be further in
the future than large-scale wind and solar.
Hydroelectric energy is the most widespread renewable energy source
in the world today, providing 16 percent of the world’s electricity. Despite
the many advantages of this energy source, it seems unlikely that this
power source can be greatly increased. Many of the world’s big rivers are
394

already dammed, and new dams often cause local environmental problems
that generate significant local political opposition.
One of the most contentious options for reducing greenhouse-gas
emissions is nuclear energy. Currently, nuclear reactors generate nearly 16
percent of the world’s electricity. Although nuclear energy is not
technically a renewable energy source, with the technology to recycle and
reprocess spent nuclear fuel, there are centuries’ worth of uranium in the
ground, even assuming a massive expansion of the world’s nuclear
generation capacity. Nuclear is a mature technology, so there is no
question about its technical feasibility.
Opponents of nuclear energy make several arguments against this
form of energy. The first is reactor safety, a problem dramatically
demonstrated by the 1986 meltdown of a reactor at Chernobyl, during
which errors by the operators caused an explosion and fire in a nuclear
reactor. The subsequent release of radioactivity to the atmosphere resulted
in an environmental disaster in the region around the reactor as well as
radioactive fallout across much of Europe. In addition, nuclear power
plants present attractive targets to terrorists, and the prospect of an attack
large enough to breach the reactor core and release its radioactive contents
to the atmosphere is truly scary.
Another problem is nuclear waste, which is what comes out of the
reactor after the nuclear fuel is burned. This waste is extraordinarily
radioactive, and it must be safely isolated for many thousands of years. If
it were released accidentally, or intentionally in a so-called dirty bomb, the
resulting harm in both human cost and ecological damage could be severe.
One way to reduce the quantity of waste is to reprocess the fuel, in which
usable isotopes of plutonium and uranium are removed and converted back
into fuel for another trip through the reactor. Even with reprocessing,
395

though, some waste must be stored for a very long time – and most people
do not want the waste to be stored near them.
This leads us to the problem of proliferation. A nuclear bomb requires
only a few kilograms of uranium or plutonium. As reactor fuel is mined,
enriched, and reprocessed, there exists the possibility that these small
amounts of bomb-grade uranium or plutonium could be diverted with the
intent of building a nuclear bomb. The diversion could occur by theft from
a legitimate nuclear program, or it could be the explicit goal of a rogue
state’s nuclear program. The net result would be a nuclear weapon in the
hands of terrorists or unstable rogue nations, which would present a
significant security threat to the rest of the world. This is why the United
States and many other countries are so opposed to Iran’s development of a
nuclear energy program.
Finally, there is the cost. Although nuclear power plants are relatively
cheap to run, they are extraordinarily expensive to build. This is one of the
primary reasons that no new nuclear power plants have been built in the
United States since the 1970s. It may be that nuclear energy cannot be
widely deployed without the government playing a key role in financing
the construction.
A final option to generate energy without emitting carbon dioxide to
the atmosphere is known as carbon capture and storage, also known by its
initials CCS, or carbon sequestration. This refers to a process by which
fossil fuels are burned in such a way that the carbon dioxide generated is
not vented to the atmosphere. Rather, the carbon dioxide is captured and
placed in long-term storage. CCS is a climate-safe technology, but is not
renewable (because you are ultimately just burning fossil fuels).
CCS is almost always used in combination with coal combustion,
because coal is abundant and produces large amounts of carbon dioxide
per joule of energy. An example of a CCS technology is to expose the coal
396

to steam and carefully controlled amounts of air or oxygen under high
temperatures and pressures. Under these conditions, atoms in coal break
apart and react with the water vapor, producing a mixture of hydrogen,
carbon dioxide, and several other gases. The carbon dioxide is separated
out, and the other gases are burned in order to generate electricity.
Once captured, the carbon dioxide must be stored. The most likely
place to store the carbon dioxide is to inject it deep underground into
porous sedimentary rocks, which are widely distributed around the world.
Particularly promising sites include depleted oil and gas fields, unminable
coal beds, or deep saline formations. This process is technically feasible
and would use many of the same technologies that have been developed by
the oil and gas industry to enhance the recovery of oil from aging fields.
The capacity of these rocks is large enough that they could conceivably
hold all of the carbon emitted by human activities.
Using available technology, approximately 85 to 95 percent of the
carbon dioxide produced can be captured. This comes at a price, however.
A power plant equipped with a CCS system would need to divert 10 to 40
percent of the energy generated into capturing and storing the carbon.
Thus, adding CCS would add a few cents per kilowatt-hour of electricity
(onto a price in the United States of 10 to 20 cents per kilowatt-hour).
The world has copious reserves of coal, and in our never-ending quest
for power, it seems likely that this coal will eventually be burned. CCS
may be the only way to simultaneously burn this coal while avoiding
climate change. However, although CCS is a promising technology, no
large power plant using CCS has ever been built, so the approach remains
unproven.
397

11.2.2 Policies to reduce carbon emissions
So the basic question of how to mitigate climate change is really a
question of how to encourage the world to switch away from fossil fuels to
carbon-free energy sources, such as wind, solar, nuclear, and CCS.
Whatever policy we adopt must be able to do this at a sufficiently low
cost, with sufficiently little social disruption and without working at cross
purposes with other societal goals, such as reducing world poverty and
conserving biodiversity, that it can maintain political support over the
decades required to address the problem.
Let us begin by addressing the question of why we need regulations
to reduce greenhouse-gas emissions in the first place. If people value a
stable climate, would not the free market take consumers’ interests into
account and reduce greenhouse-gas emissions without any government
intervention? There is, after all, abundant evidence that the free market is
efficient at allocating resources and producing socially beneficial
outcomes (although, as the occasional economic meltdown shows, it is not
perfect). As U.S. Senator Chuck Hagel said in an interview,2
I have always believed that the marketplace does work. It works
because it’s based on one fundamental dynamic, which is self-interest
of an individual, a company, or a country. The marketplace fosters
competition and always trends toward producing a better, cheaper
product, which means it is a driver of efficiency. It’s in the interests of
everyone here to make a cheaper product that’s less energy intensive.
It cleans up the environment, which has economic advantages too.
So if switching to carbon-free energy is something we want to do, would
not the free market take care of this all by itself? The answer is no –
398

emissions reductions sufficient to stabilize the climate will not occur by
themselves.
There is, of course, some truth to Hagel’s argument. Because
consumers have to pay for electricity, consumers have an incentive to
reduce how much electricity they consume. This applies pressure to
manufacturers to design, build, and sell equipment that uses as little energy
as possible. As a result of this market pressure, just about every piece of
equipment that you buy today is more efficient than the comparable piece
of equipment that was available a few decades ago. That is one of the
primary factors behind the world economy’s long-term decrease in energy
intensity discussed in Section 8.2.
But it is also important to remember that increases in energy
efficiency are, by themselves, insufficient to solve the climate change
problem. The bulk of emissions reductions will come from a large-scale
shift toward carbon-free energy that is not currently occurring. Moreover,
there is a good reason it is not now occurring – and may never occur
without government intervention.
To understand why, consider the following scenario. Imagine you
own a company that produces widgets. Let us assume that it costs your
company $1 to manufacture a widget, and during the manufacture process,
some carbon dioxide is released into the atmosphere. This carbon dioxide
will cause climate change, which causes damages valued at 10 cents. Thus,
the total cost of manufacturing a widget using this process is $1.10.
However – and this is important – the widget manufacturer only pays $1;
the costs of climate impacts associated with the widget, 10 cents, are borne
by everyone in the world. At the same time, the benefit of producing the
widget goes entirely to the manufacturer.
Economists call the costs of climate change imposed on the rest of the
world by the widget manufacturer an externality. More generally, an
399

externality occurs when someone takes an action, and this action imposes
involuntary costs on others. To understand the economic implications of
an externality, imagine that someone invents a new process for
manufacturing widgets, in which it costs $1.05 to produce a widget but no
carbon dioxide is emitted to the atmosphere. There are no climate impacts
associated with this process, so the total cost is also $1.05 – and this cost is
entirely borne by the manufacturer. Because the total cost of building a
widget under the new process is 5 cents cheaper than the total cost for the
older process, it would be beneficial to society for the widget manufacturer
to switch to this new process.
But the widget manufacturer will not switch. The widget
manufacturer is only paying $1 per widget under the older method, with
the rest of the cost being borne by society. Under the new process, the
manufacturer pays the entire cost of $1.05. So even though this new
process is cheaper to society as a whole, to the widget manufacturer, this
new process is more expensive. Thus, because of the externality, the
socially and economically preferred outcome does not occur, a result
sometimes referred to as a market failure.
Externalities and the associated market failures occur frequently in
environmental problems in which some profitable economic activity
degrades a common asset such as the atmosphere, the ocean, or a river –
and the costs of that degradation are paid for by everyone. In such a
situation, the incentive is to utilize the asset without regard to its
degradation because those costs are paid by everyone in the society – not
the person actually damaging the asset. This is a version of the famous
“tragedy of the commons,” and it is the fundamental economic explanation
for why overfishing is depleting stocks of fish in the oceans, logging is
destroying the rainforests, and greenhouse-gas emissions are changing the
climate.
400

In the case of greenhouse gases, emitters exploit the atmosphere by
dumping carbon dioxide into it. Because it is free to load the atmosphere
with carbon dioxide, the rational behavior of each emitter is to dump as
much greenhouse gas into the atmosphere as is necessary to maximize
profit. And there is no incentive to reduce emissions. If a company decides
to reduce emissions, the entire cost of reducing emissions is borne by the
company – while the benefits are spread throughout the society.
Thus, the root economic cause of the climate change problem is that it
is free for emitters to dump greenhouse gases into the atmosphere.
Therefore, to solve climate change, most economists argue that the costs
associated with the emission of greenhouse gases must be shifted back
onto the emitter. In the parlance of economics, we need to internalize the
externality. This is also frequently described by the principle of “polluter
pays.” meaning that emitters should be held accountable for the damage
they cause.
If the widget manufacturer had to pay 10 cents in order to emit the
carbon dioxide associated with the production of a widget under the old
manufacturing technology, the manufacturer would be paying the entire
cost of manufacturing the widget. In that case, it would be cheaper for the
manufacturer to switch to the new method, which costs $1.05 to produce a
widget with no emissions of greenhouse gases. This benefits both the
manufacturer and society.
The bottom line on mitigation: Most experts view mitigation efforts
as a necessity. To avoid warming that is generally considered dangerous,
emissions have to be reduced below present levels by 50 to 80 percent by
the middle of this century.
Although there are many ways to reduce emissions, the only practical
way is through technology. Improvements in energy efficiency are
important – and make sense even if climate change is not an important
401

problem – but such changes cannot by themselves produce the deep
reductions in emissions required. Rather, changes in energy-generation
technology are going to be required. This means switching to technologies
that generate energy without emitting greenhouse gases, which include
solar, wind, biomass, nuclear, and carbon capture and storage.
To encourage the adoption of new technologies, most proposed
mitigation policies put a price on emissions. Right now, emitting carbon
dioxide to the atmosphere is free, and the costs of the resulting climate
change are imposed on everyone in the world. In this situation, there is no
incentive for the emitter to reduce emissions of greenhouse gases. Putting
a price on emissions makes emitters pay the full cost of their emissions,
thereby giving them an incentive to adopt climate-safe energy technology.
In Chapter 12, I will discuss pricing carbon in great detail.
402

11.3 Geoengineering
A last solution to the climate change problem is known as geoengineering,
which refers to actively manipulating the climate system in order to
prevent the climate from changing – or even to return the climate to some
past condition. The idea is that we would continue burning fossil fuels and
emitting the resulting carbon dioxide into the atmosphere, but we would
also make other changes that tend to cool the climate, offsetting the
warming from the emissions.
Geoengineering efforts can be roughly divided into two categories:
solar radiation management and carbon-cycle engineering. Solar radiation
management schemes engineer a reduction in the amount of solar energy
absorbed by the Earth, Ein = S (1 ‒ α)⁄4, where S is the solar constant and α
is the albedo (if you do not remember this, you should review Chapter 4).
Most solar radiation management schemes cool the Earth by
increasing the albedo. Probably the most frequently discussed way to do
this is to inject sulfur dioxide (SO2) into the stratosphere. Once in the
stratosphere, this gas reacts with water vapor to form aerosols – liquid
droplets that are so small that they have negligible fall speed (discussed in
Chapter 6). These aerosols reflect sunlight back to space, thereby
increasing the albedo of the Earth and leading to cooling. Injection of
sulfur into the stratosphere is the same mechanism by which volcanoes
cool the planet.
Another option is to increase the reflectivity of clouds. It turns out
that the size of the cloud droplets determines how white a cloud is, with
smaller cloud droplets making a cloud whiter or more reflective. This is
the same reason that powdered sugar, which is made up of small particles,
403

appears whiter than chemically identical table sugar. Thus, if we could
somehow make the particles in clouds smaller, clouds would become more
reflective and raise the albedo of the Earth.
One way to do this is to release what are known as cloud
condensation nuclei into the cloud. These nuclei serve as seeds that cloud
droplets form around. By adding them to clouds, we increase the total
number of droplets in the cloud. Because the total water contained in a
cloud is basically fixed, this makes the cloud contain more, but smaller,
particles. This effect can be seen in ship tracks in the clouds, which were
discussed in Chapter 6. As ships steam across the ocean, the exhaust from
their diesel engines contains fine particulates that can serve as cloud
condensation nuclei. These aerosols are transported by the winds into low-
level clouds and brighten them.
The physics supporting these suggestions is robust, and we have high
confidence that if any of these schemes were carried out at sufficiently
large scale, the planet would indeed cool. There are, however, important
disadvantages with these approaches. The first is that solar radiation
management schemes focus on temperature, but temperature increases are
only one of many impacts associated with climate change – and perhaps
not even the most important. We know, for example, that some of the
carbon dioxide released into the atmosphere ends up in the ocean, resulting
in ocean acidification. Solar radiation management schemes do nothing to
address this impact of continued carbon dioxide emissions.
Moreover, solar radiation management schemes may create other
problems. The 1991 eruption of Mount Pinatubo, for example, led to
substantial changes in global precipitation patterns and an increase in the
incidence of drought in some regions. We can therefore expect that
reducing the amount of solar radiation reaching the Earth, in addition to
cooling the planet, would also lead to changes in the amount and
404

distribution of global precipitation. Whether these changes would be better
or worse than climate change is a question whose answer is not clear.
And there are important political problems with this approach.
Imagine that a few rich countries in the world (e.g., the United States and
Europe) got together to inject sulfur into the stratosphere to cool the
planet. Then, China or India experienced a severe drought. Whether the
sulfur injection caused that drought or not, the countries affected might
well believe that it did. It is easy to imagine that this would lead to a great
amount of political tension, possibly even the abandonment of the
geoengineering effort. In the worst case, geoengineering by a group of
countries might be considered an act of war by another group of countries
that suffer some type of weather-related injury at the same time.
In addition, as long as our society is increasing the amount of
greenhouse gas in the atmosphere, geoengineering efforts must be
continually strengthened in order to provide an ever-increasing cooling
influence to keep the climate stable. So if we were, for example, injecting
sulfur into the stratosphere in order to enhance the planet’s albedo, then we
would have to be injecting ever greater amounts of sulfur over time in
order to offset a continually increasing abundance of atmospheric
greenhouse gases.
If we ever stopped injecting sulfur into the stratosphere, it would take
only a few years for the stratosphere to clear (about the same length of
time as it does for the climate to recover after a volcano), during which
time the Earth’s albedo would be rapidly decreasing. This would cause a
significant rise in the Earth’s temperature over the following few decades –
which would be very bad. Thus, once you start geoengineering, it is
difficult to stop.
The second category of geoengineering, carbon-cycle engineering,
modifies the carbon cycle so that carbon dioxide is more rapidly removed
405

from the atmosphere. As we learned in Chapter 5, it takes a few centuries
for the majority of carbon dioxide emissions to be removed from the
atmosphere (complete removal takes much longer), and this is the reason
carbon dioxide is such a pernicious greenhouse gas. If the lifetime of
carbon dioxide could be reduced, it would reduce the amount of carbon
dioxide in the atmosphere, thereby reducing the amount of climate change.
Planting trees is an example of carbon-cycle engineering. As the trees
grow, they suck carbon dioxide out of the air and sequester it in wood.
Another scheme is to add iron to the ocean. Iron is thought to be a limiting
nutrient there, so the addition of iron will stimulate the growth of
phytoplankton. As the phytoplankton grow, carbon dioxide will be drawn
out of the atmosphere and into the ocean. The phytoplankton are then
consumed by larger organisms, and subsequent biological activity creates a
rain of dead organisms and fecal matter from surface waters into the deep
ocean. Thus, adding iron to the ocean has the net effect of drawing carbon
dioxide out of the atmosphere and transporting it to the deep ocean.
Another option is to remove carbon dioxide from the air chemically,
which is often referred to as air capture. This is like CCS, but CCS
removes carbon dioxide from the hot exhaust gas of a power plant whereas
air capture removes carbon from the free atmosphere. This is an attractive
option, but the amount of energy required is staggering, and this severely
limits our ability to deploy this technology at scales large enough to
remove significant quantities of carbon dioxide from the atmosphere.
In general, carbon-cycle engineering is attractive because, unlike
solar radiation management, it does not focus just on temperature. If we
balance emissions of carbon dioxide from human activities with removal
through carbon-cycle engineering, we will truly stabilize the climate – not
just temperature but other aspects, such as ocean pH and precipitation. In
fact, a sufficiently aggressive program could lead to a reduction in
406

atmospheric carbon dioxide (not just a stabilization), which would
eventually undo most of the effects of climate change (but not all – some
changes, such as extinction of species, are completely irreversible, whereas
others, such as loss of the world’s largest ice sheets, are effectively
irreversible on any time scale that we care about).
There are, of course, some problems with the various approaches to
carbon-cycle management. For example, real engineering of the carbon
cycle, such as adding iron to the ocean, is risky. Because of significant
uncertainties in our knowledge of how carbon cycles in the ocean, there is
a possibility that that these schemes will not work. In other words, we
might add iron to the ocean only to find out that, because of unanticipated
physics or biology, no extra carbon dioxide was removed from the
atmosphere. Even worse, it might have unforeseen and serious impacts on
ocean ecosystems. Thus, in trying to address climate change, we may
cause an entirely new environmental problem – and not even solve the
problem we were intending to address.
Overall, geoengineering has several general qualities that make it an
attractive policy response to climate change. First, and possibly most
importantly, it attacks the climate change problem through technology but
does not seek to limit emissions in any way. That allows us to continue
doing exactly what we are doing now: burning fossil fuels and consuming
energy as fast as we possibly can. Second, geoengineering is a relatively
rapid response. Many schemes can be implemented in a decade or so. This
contrasts favorably with mitigation, for which emissions reductions must
begin now in order to head off climate changes that will occur in 50 to 100
years. Thus, like adaptation, geoengineering is often thought of as
something that we will do in the future, if climate change turns out to be
bad.
407

Geoengineering also avoids many of the moral problems of
adaptation. A successful geoengineering effort will help the entire planet,
not just the rich countries. Moreover, the cost estimates are often cheap
enough that one or a few rich countries could shoulder the burden.
However, it also raises moral problems of its own: Should we be
engineering the climate? Who decides what the ideal temperature is? What
happens if geoengineering efforts harm one particular region while
benefiting all others? These are all difficult questions that have to be
resolved before the world uses geoengineering to address climate change.
The bottom line on geoengineering: Geoengineering is an appealing
but risky approach to dealing with climate change. Although it may work,
the risk exists that geoengineering may lead to unintended consequences
that would leave the world worse off. And some approaches do not address
all of the impacts of climate change and may only be viable for a relatively
short period of time (less than a century). Thus, in a world where climate
change is handled responsibly through mitigation and adaptation, it is
unlikely that geoengineering would be needed.
Nevertheless, it is easy to imagine a future in which the world makes
no progress in reducing emissions. If, by the middle of the twenty-first
century, emissions are large and climate change is out of control,
geoengineering may represent the last hope for avoiding truly disastrous
climate change. It is in this type of scenario that the deployment of
geoengineering is a reasonable strategy. But even here, geoengineering is
unlikely to be the final solution. Rather, it will be a way to buy time to let
emergency mitigation efforts reduce emissions.
408

11.4 Chapter summary
Responses to climate change can be roughly divided into three
categories: adaptation, mitigation, and geoengineering.
Adaptation means learning to live with climate change; mitigation
refers to reducing emissions of greenhouse gases, thereby
preventing the climate from changing in the first place; and
geoengineering refers to active manipulation of the climate system
in order to engineer a cooler climate.
Because not all climate change can be prevented, we will have to
adapt to some climate change. However, there are both moral and
practical problems with relying on adaptation as our only response.
Mitigation refers to efforts to reduce emissions, thereby preventing
future climate change. Most experts and world leaders view
mitigation as a vital component of any plan to address climate
change. This will be accomplished mainly by transitioning from
fossil fuels to energy sources that do not emit greenhouse gases. A
key part of policies designed to encourage transition to climate-safe
energy is to put a price on emissions of greenhouse gases, which
are presently free.
Geoengineering is an active manipulation of the climate system.
Under this approach, our society would continue adding
greenhouse gases to the atmosphere, but we would intentionally
change some other aspect of the climate in order to cancel the
warming effects of the greenhouse gases. Geoengineering
strategies can be broken into two categories – solar radiation
409

management and carbon cycle engineering. Because of potential
problems with geoengineering, it is generally considered a last-
ditch approach, only to be used if adaptation and mitigation
approaches fail.
410

Additional reading
Much has been written about our policy response to climate change in the
form of books, reports, and scholarly publications. Many are available
online, and you can find them by searching the Internet. Here are a few of
particular note.
The IPCC’s Working Group II covers impacts and adaptation, and it
remains one of the most authoritative summaries of what we know. You
can download and read the most recent report at ipcc-
wg2.gov/AR5/report/. The IPCC’s Working Group III focuses on
mitigation of climate change. You can download the most recent report
from this working group at mitigation2014.org/.
For a more U.S.-centric view of adaptation and mitigation, see these
reports: National Research Council, Adapting to the Impacts of Climate
Change (Washington, DC: The National Academies Press, 2010);
available online at www.nap.edu/catalog.php?record_id=12783 and
National Research Council, Limiting the Magnitude of Future Climate
Change (Washington, DC: The National Academies Press, 2010);
www.nap.edu/openbook.php?record_id=12785.
The key to understanding why we have a climate problem is recognizing
that there are “hidden” costs to energy that are not paid by the consumers,
known as externalities. This comprehensive report details these
surprisingly hefty costs (National Research Council, Hidden Costs of
Energy: Unpriced Consequences of Energy Production and Use
[Washington, DC: The National Academies Press, 2010]; available online
at www.nap.edu/catalog.php?record_id=12794).
411

http://ipcc-wg2.gov/AR5/report/

http://mitigation2014.org/

http://www.nap.edu/catalog.php?record_id=12783

http://www.nap.edu/openbook.php?record_id=12785

http://www.nap.edu/catalog.php?record_id=12794

D. J. C. MacKay, Sustainable Energy – Without the Hot Air (Cambridge:
UIT Cambridge, 2009). This is a sober and quantitative analysis of how
hard it will be to replace fossil fuel energy with carbon-safe energy
sources. The short answer is that it would not be easy (download the book
at www.withouthotair.com).
There are several good and very readable books about geoengineering out
there. These include E. Kintisch, Hack the Planet: Science’s Best Hope –
or Worst Nightmare-for Averting Climate Catastrophe (New York: Wiley,
2010) and J. R. Fleming, Fixing the Sky: The Checkered History of
Weather and Climate Control (New York: Columbia University Press,
2012).
See www.andrewdessler.com/chapter11 for additional resources for this
chapter.
412

http://www.withouthotair.com

http://www.andrewdessler.com/chapter11

Terms
Adaptation
Air capture
Carbon capture and storage
Carbon-cycle engineering
Carbon-free-climate-safe energy sources
Carbon sequestration
Externality
Geoengineering
Market failure
Mitigation
Renewable energy
Solar photovoltaic
Solar radiation management
Solar thermal
413

Problems
1. Our responses to climate change can be put into three general
categories. List the categories. For each category, give one example
of an action that would fall into that category.
2.
a) What are carbon-free energy sources? List the ones discussed in
the book.
b) Is carbon-free energy the same as renewable energy?
c) Is nuclear energy carbon free? Is it renewable?
3. Why do economists generally believe that the free market will not
solve the climate problem by itself?
4. What is an externality?
5. Do some research and find an example of “the tragedy of the
commons.” Explain how global warming is an example of this type of
problem?
6.
a) Your friend says that “we should rely entirely on adaptation as
our response to climate change.” Is this a good idea?
b) I argued here that adaptation must be at least part of our
response. Why?
7.
a) Explain one way we can “geoengineer” a higher planetary
albedo.
414

b) Explain one way we can “geoengineer” a reduction in carbon
dioxide.
8.
a) In this chapter, we divided geoengineering approaches into two
categories. What are they?
b) What are the advantages and disadvantages of geoengineering?
c) I argued that that geoengineering should be used in what
circumstances?
9.
a) Imagine a credit card whose bill was divided up and sent to
everyone in the United States (i.e., if you purchased something on
this card, every person in the United States would get a bill for
1/300,000,000th of your total cost). Would the average person
spend freely with this credit card? Or would they be as thrifty as
they would if they had to pay the entire bill?
b) Now imagine that every person in the United States has a credit
card like this. What do you think is going to happen?
c) How is this situation related to the climate change problem?
10. In this chapter, we explored the terms of the IPAT relation and
concluded that reducing emissions could really be achieved only
through reduction of one term. Which term is it, and why is that our
only real option?
11. The technology term in the IPAT relation can be further divided
into two terms.
415

a) For each term, give an example of a technological switch that
reduces it.
b) One of these terms is the key to deep reductions in emissions.
Which one is it? What kinds of changes are required to make such
deep reductions?
12. Why do mitigation policies have little ability to influence the
climate over the first half of the twenty-first century?
13. In this chapter, we talked about negative externalities. Can you
think of an example of a positive externaility?
1 By encouraging the undesirable outcome, this is an example of what is
often referred to as a “moral hazard.”
2 www.grist.org/article/hagel/.
416

http://www.grist.org/article/hagel/

12
Mitigation policies
◈
Chapter 11 discussed the three options we have to address climate change:
adaptation, mitigation, and geoengineering. Adaptation will, by necessity,
be an important part of our response to climate change. However, relying
entirely on adaptation as our only response to climate change is fraught
with problems. Geoengineering is another possibility, but one that few
people think should be used now. Rather, it is the last resort – like an
airbag in a car – that you turn to if other approaches to address climate
change fail.
The remaining option is mitigation – the reduction of greenhouse
gases emissions so as to avoid climate change – and there is agreement
among those who have looked seriously at the problem that we should
embark on mitigation efforts right now. Mitigation schemes will have little
effect on the climate of the next few decades, but a successful mitigation
effort would allow us to avoid large climate changes in the second half of
this century and beyond.
As we learned in Chapter 11, reducing greenhouse-gas emissions
requires improvements in the energy efficiency of our economy and, most
importantly, converting our energy system to one that primarily utilizes
carbon-free energy sources, such as solar, wind, nuclear, and CCS. We
417

also discussed in the last chapter why these emissions reductions would
not occur without explicit government policy. In this chapter, we explore
in detail the policy options that governments can use to reduce emissions.
418

12.1 Conventional regulations
The conventional approach to regulation, often described colloquially as
command-and-control regulation, requires all emitters in a particular
economic sector to meet a single standard. Electricity companies, for
example, might be required to generate energy by using a particular
technology, such as wind or CCS. Cars might be required to be gas/electric
hybrids. Alternatively, regulations may limit total emissions of a pollutant,
or enforce a standard of greenhouse gases emitted per kilowatt-hour
generated (for power plants) or greenhouse gases emitted per mile driven
(for cars). For example, in 2012 the Obama Administration proposed
requiring new power plants to emit less than 1,000 pounds of carbon
dioxide per megawatt-hour of energy produced. Burning coal produces
more than this, so this regulation effectively prohibits the construction of
new coal-fired power plants (without CCS).1
The conventional approach has the advantage that it is clear and easy
to understand. Even today, many environmental regulations, including
regulations on air pollution, fall into this category. Since the 1980s,
however, weaknesses with this approach have been identified, and it has
been falling out of favor with regulators. First, technologies specified (e.g.,
wind, CCS) may not actually turn out to be the best ones. Second, the
regulations force all emitters to meet the same emissions standards. This
ignores the fact that some emitters can reduce emissions more cheaply
than others. Third, conventional regulations provide no incentive for the
development and deployment of new technologies that reduce emissions
beyond the specified target.
419

Because of these disadvantages, there is little talk in policy circles of
attempting to solve the climate change problem primarily with
conventional regulations. Instead, market-based regulations are now
preferred, and I will spend the rest of the chapter discussing them.
420

12.2 Market-based regulations
In Chapter 11, I discussed why the free market is unlikely to solve climate
change without intervention from the government: The basic reason is that
it is free to dump greenhouse gases into the atmosphere. The cost of these
emissions is imposed on everyone, not the emitter — which is why
economists call this an externality. Because the costs of the emissions are
not borne by the emitter, there is no economic incentive for the emitter to
reduce emissions. The solution therefore is to make emitters pay for
emitting greenhouse gases. In so doing, we give the emitters an economic
incentive to reduce emissions.
Making emitters pay for their emissions is a market-based solution. It
does not tell anyone how much they can emit or what technology to use –
it only requires them to pay for whatever emissions they do make. And it
makes reductions more cheaply than convectional regulations. In the next
two sections, I will discuss the two market-based regulatory approaches
most frequently discussed in the climate change policy debate: carbon
taxes and cap-and-trade systems.
421

12.2.1 Carbon tax
The first approach is a carbon tax. Under this policy, emitters have
complete freedom to emit as many tons of greenhouse gas to the
atmosphere as they choose, as long as they pay a specified fee to the
government for each ton released to the atmosphere.
To understand why a carbon tax reduces emissions, let us imagine a
power plant, which we will call Plant A, that emits 10 tons of carbon
dioxide into the atmosphere each year. The third column of Table 12.1
contains the marginal cost for reducing emissions for Plant A, which is the
cost of reducing a particular ton of emissions. Thus, Plant A can reduce its
annual emissions by 1 ton – from 10 tons to 9 tons – for $1. Reducing
emissions another ton, from 9 tons to 8 tons, costs an additional $2. Thus,
the total cost of reducing emissions from 10 to 8 tons is the sum of the
marginal costs,2 that is, $1 + $2, for a total cost of $3. The next ton of
emissions costs $3 to eliminate, so the total cost of reducing emissions
from 10 to 7 tons is $1 + $2 + $3 = $6. And so on.
Table 12.1 Cost of reducing emissions for Plants A and B
Emissions
reduced by
(tons)
Units
emitted
(tons)
Plant A’s cost ($) Plant B’s cost ($)
Marginal Total Marginal Total
0 10 – – – –
1 9 1 1 2 2
2 8 2 3 4 6
3 7 3 6 6 12
422

4 6 4 10 8 20
5 5 5 15 10 30
6 4 6 21 12 42
7 3 7 28 14 56
The marginal cost of reducing emissions increases as emissions are
progressively reduced. To see why, think about golf. When you just start
out, you may be shooting 130 strokes over eighteen holes. It takes
relatively little effort to reduce your score by one stroke, down to 129 –
perhaps only 1 hour at the driving range. Taking another stroke off your
score, reducing your score to 128, takes slightly more work, maybe two
more hours at the driving range. By the time your score reaches 80 and
you are getting close to par, it can be extremely difficult and take
enormous practice to reduce your score by one stroke to 79. Eventually, as
you approach and possibly surpass par, you reach a limit beyond which
you will never move past, no matter how hard you work.
In this golf example, the marginal cost is the amount of effort you
have to apply to reduce your golf score by one stroke. The fact that the
amount of work it takes to improve your score by one stroke increases as
you get better is a version of the law of diminishing returns: each
additional unit of effort (e.g., an hour at the driving range) produces less of
an improvement than the previous unit of effort.
For Plant A, fine-tuning the machinery in the plant may reduce the
first ton at very little cost. Once the equipment has been tuned up,
however, additional reductions are harder to make. Eliminating the second
ton may require replacing some outdated equipment with newer, more
efficient equipment. This will cost more than the first ton. The third ton
may require even more equipment replacement, or perhaps wholesale
423

changes in the plant’s operation. This ton will therefore be more expensive
to eliminate than the previous two tons.
Now imagine that a carbon tax of $4 per ton is imposed on the
emitters – meaning that for every ton that is emitted into the atmosphere,
the plants have to pay the government $4. How would each plant respond?
Remember that Plant A has total freedom to emit as much as it wants – the
carbon tax does not specify any reduction. Plant A will therefore search for
its cheapest alternative.
Plant A can emit the tenth ton and pay a tax of $4, or it can pay $1
and not emit that ton. It does not take a financial genius to conclude that
the rational thing to do is to not emit the ton. Now emissions are down to 9
tons, and Plant A can emit the ninth ton and pay a tax of $4, or it can pay
$2 and not emit that ton. For this ton, too, the rational thing to do is to not
emit that ton. Now emissions are down to 8 tons, and Plant A can emit the
eighth ton and pay a tax of $4, or it can pay $3 and not emit that ton.
Again, the rational thing to do is to not emit that ton.
Now things get a bit trickier. For the 7th ton, Plant A can emit the ton
and pay a tax of $4 or it can pay $4 and not emit that ton. From a purely
financial point of view, these two alternatives are equivalent. I suspect,
though, that most companies will reduce that last unit, because it is the
same cost as the tax and the company can use this to burnish its
environmental reputation. So we can assume that Plant A will choose to
not emit the seventh ton.
For the sixth ton, Plant A can emit the ton and pay a tax of $4 or it
can pay $5 and not emit that ton. The rational thing to do in this situation
is to pay the tax and emit that ton. Thus, under a carbon tax of $4, Plant A
will reduce emissions by 4 tons.
Now let us consider a second plant, which we will call Plant B, whose
marginal costs are also listed in Table 12.1. Plant B can emit the tenth ton
424

and pay a tax of $4, or it can pay $2 and not emit that ton. Clearly, Plant B
will not emit the ton. Now emissions are down to 9 tons, and Plant B can
emit the ninth ton and pay a tax of $4, or it can pay $4 and not emit that
ton. As for Plant A, we can assume that Plant B will choose to not emit
that ton. Now emissions are down to 8 tons, and Plant B can emit the
eighth ton and pay a tax of $4, or it can pay $6 and not emit that ton. Here,
the rational thing to do is to pay the tax and emit that ton. Thus, under a
carbon tax of $4, Plant B will reduce emissions by 2 tons.
Thus, Plant B reduces emissions less than Plant A does for the same
carbon tax rate. The reason is that Plant B has higher marginal costs for
reducing emissions. This may arise for any number of reasons – for
example, Plant B may be older than Plant A and so be using outdated
technology that is not amenable to reducing emissions.
To summarize, under a carbon tax each emitter will reduce emissions
until the marginal cost of reduction is equal to the carbon tax rate. For
Plant A, the marginal cost equals the tax rate of $4 per ton when emissions
have been reduced 4 tons, whereas for Plant B, the marginal cost equals
the tax rate when emissions have been reduced by 2 tons. Because
marginal costs vary among emitters, some emitters will make deeper cuts
than others.
The total reduction in emissions from the two plants in response to a
carbon tax of $4 per ton is 6 tons: a reduction of 4 tons from Plant A and 2
tons from Plant B. The total cost to Plant A of reducing emissions is $1 +
$2 + $3 + $4 = $10, whereas the total cost to Plant B is $2 + $4 = $6.
Thus, the total cost to society of reducing 6 tons of emissions is $10 + $6 =
$16.
Under a conventional command-and-control approach, there is a
single performance target that each plant is required to meet. So an
emissions reduction of 6 tons might be achieved, for example, by having
425

both Plant A and B reduce emissions by three tons. This would cost Plant
A $1 + $2 + $3 = $6 to reduce 3 tons, whereas it would cost Plant B $2 +
$4 + $6 = $12 to reduce 3 tons. The total cost of this 6-ton reduction is
$18.
This is an important result. The carbon tax of $4 per ton resulted in a
6-ton reduction for $16, whereas the conventional command-and-control
approach resulted in a 6-ton reduction for a cost of $18. The carbon tax is
cheaper because of its flexibility – it shifts reductions to the lowest
marginal cost emitters, in this case, Plant A – so that the emissions
reductions are made where they are cheapest, which lowers overall cost to
society.
A carbon tax would be reasonably easy to implement. Most
greenhouse gases come from fossil fuels, and these are produced at a
relatively small number of sites. A carbon tax could be applied to the fossil
fuel when it is extracted from the ground, using the administrative
infrastructure for existing taxes, such as excise taxes on coal and
petroleum. The price of the tax would then follow the fuel through the
market, where the end user would finally pay it. A tax credit would be
generated if the carbon is used in such a way that it was not released into
the atmosphere (such as production of plastic or capture of carbon in coal
combustion followed by sequestration).
As part of a long-term policy, the carbon tax would start out relatively
small and, over several decades, gradually increase until emissions have
been reduced to the target level. Gases other than carbon dioxide, such as
methane or nitrous oxide, would also be taxed but at a rate that takes into
account how effective each one is at warming the planet. For example, 1
ton of methane contributes approximately twenty times more warming
than a ton of carbon dioxide, so the tax on methane should be
proportionately higher than the tax on carbon dioxide.
426

The costs of reducing emissions would eventually be passed on to
consumers. Thus, the net effect of a carbon tax is to raise the prices of
goods and services by an amount proportional to the amount of the
greenhouse gases released. Goods and services that are produced with little
or no emission of greenhouse gases will not experience price increases,
whereas the costs of goods and services that require the emission of
significant amounts of greenhouse gases may see large price increases.
Many people automatically consider taxes to be bad, so they look at
suggestions of a carbon tax with, to put it mildly, disdain. However, most
economists argue that a well-designed tax serves a useful economic
purpose. For activities that generate negative externalities (costs imposed
on society, such as emitting greenhouse gases or smoking cigarettes), the
free market prices these activities too low, leading to overconsumption of
the associated good or service. Taxes on these activities correct for this
and reduce consumption, which produces a socially beneficial outcome.
Thus, an economist thinks of a carbon tax as fixing a problem in the free
market. Unfortunately, given the automatic opposition that new taxes
generate, even those that make economic sense, the present prospects for
implementing a carbon tax in the United States and in many other
countries are dim.
427

12.2.2 Cap and trade
An alternative way to put a price on greenhouse-gas emissions is a cap-
and-trade system. Under cap and trade, the government issues a fixed
number of permits each year, with each permit allowing the holder to emit
a fixed amount (often 1 ton) of greenhouse gas to the atmosphere. Emitters
must hold permits for the amount of greenhouse gas they emit to the
atmosphere. Thus, the total number of permits issued sets a cap on total
emissions. Emitters with extra permits can sell them to those needing
additional permits (hence the trade part). The price of the permits is set by
the market, not by the government.
The economics of a cap-and-trade system is nearly identical to the
economics of the carbon tax. If the marginal cost of reducing 1 ton of
greenhouse gas emissions is less than the cost of the permit, the emitters
will not emit that ton. This allows them to either avoid having to buy a
permit, or, if they already have a permit for this ton, they can sell it at a
profit. If the marginal cost is more than the permit, the emitters will
acquire a permit and emit that ton. In the end, the emitters will reduce
emissions until the marginal cost of reducing emissions equals the price of
the permits.
So if permits cost $4 per ton, Plant A will use permits to emit 6 tons,
thereby reducing emissions by 4 tons. Plant B will use permits to emit 8
tons, thereby reducing emissions by 2 tons. This is the same result that was
obtained for a carbon tax of $4 per ton.
Under most cap-and-trade systems, when a unit of greenhouse gas is
emitted, a permit is retired. Therefore, the government must continually
issue new permits to replace those that have been used. Over several
428

decades, the number of permits issued each year will decrease following a
prescribed schedule until the target emissions level is reached.
One of the most contentious issues in any cap-and-trade system is
how the government issues those permits. One approach is for the
government to auction the permits off. In that case, companies would buy
the permits from the government and then pass the cost of the permits on
to their customers through higher prices for their products. This approach
has the advantage that permits go to those emitters who value the permits
the most – and are therefore willing to pay the most. These will be the
highest marginal cost emitters, for whom emissions reductions are most
expensive.
Emitters, like utilities that burn coal, oppose auctioning the permits
because they would have to pay the cost of the permits, which would then
be passed on to consumers in the form of higher prices. This will reduce
demand for their product – which would cost them money. The alternative
is for the government to give away permits to companies for free. Because
permits can be sold, this is equivalent to the government giving emitters
money. Unsurprisingly, emitters favor this approach.
When giving the permits away for free, the decision about how to
allocate permits is not determined by the market, like they would in an
auction, but by other issues, such as fairness and political connections. For
example, the imposition of a cap-and-trade system will be potentially
disruptive to industries that emit a lot of carbon to the atmosphere (e.g.,
coal companies). Giving these industries free permits essentially provides
financial aid to help them adjust to a new world in which emitting carbon
to the atmosphere is no longer free. Politicians can also distribute permits
to curry favor from particular constituents or to buy support for the policy
from particular industries.
429

In the most recent bill considered by the U.S. Congress in 2009 and
2010, for example, the majority of permits would initially be given away
for free. Many of these free permits would go to many large emitters,
including coal companies. Because of that, many of these large emitters
strongly favored passage of the bill. The bill then required a slow
transition to auctioning 100 percent of the permits over the next two
decades.
One last issue is that both a carbon tax and a cap-and-trade system
(with auctioned permits) would create an enormous transfer of wealth from
consumers to the government. Some of this wealth could be used to help
those with low incomes, who would be disproportionately hurt by the rise
in energy prices. The government could also use the income for other
beneficial activities, such as research and development of new energy
technology or reduction of the deficit. One frequently made suggestion is
to use this money to reduce other taxes, such as those on labor and capital,
or rebate the money evenly to every citizen (sometimes called “cap and
dividend”).
430

12.2.3 Carbon tax versus cap and trade
Carbon tax and cap-and-trade systems are quite similar in many ways.
Both reduce emissions by putting a price on emissions. Both systems allow
companies to emit as much as they want, as long as they pay the tax or
possess a permit for each unit emitted. In doing so, both move emissions
reductions to where they are cheapest, namely to the lowest marginal cost
emitters. In both cases, the emitter reduces emissions until the marginal
cost of reducing the next ton of emissions is equal to the price on
emissions.
Putting a price on emissions means that both approaches raise the
price of fossil fuels and the goods and services made from them in
proportion to the amount of greenhouse gases emitted by their
consumption. Although consumers may not like to see prices go up,
economics tells us it is the most efficient way to reach a socially optimum
level of emissions, and it does so through several mechanisms. First,
higher prices encourage consumers to reduce their consumption of
greenhouse-gas-intensive goods and services. Second, putting a price on
emissions encourages the economy to substitute climate-safe technology
for their present technology. This occurs because these policies raise the
prices of emitting greenhouse gases, so they increase the economic
competitiveness of climate-safe technologies (e.g., solar, wind, nuclear).
Third, and most importantly, it encourages research on and development of
new technologies that can replace today’s technologies that produce
greenhouse gases. Humans are amazingly clever, and putting a price on
emissions signals to the market that innovation and breakthrough
technologies will pay off handsomely. With this market incentive, we can
431

expect new technologies that could dramatically reduce the cost of
stabilizing the climate.
However, there are some important differences between these two
approaches. Under a carbon tax, the policymakers set the tax rate, which in
turn sets the cost to society of the emissions reductions. But it is not
exactly known what the economy’s marginal cost of reduction is, so this
means there is uncertainty in exactly how much of an emissions reduction
will occur given a particular tax rate. Under a cap-and-trade system, in
contrast, the policymakers set the total number of permits issued, and
therefore the total emissions from the economy. However, the uncertainty
in the marginal costs means that it is unknown how much it will cost to
reach the specified level of emissions.
Here is a simple analogy. Imagine you go to a store to buy some soda.
You are given $50 and instructions to buy as much soda as you can. In that
case, you know the total cost you will incur upfront ($50), but you do not
know how much soda this will purchase. This is analogous to a carbon tax.
This creates the potential problem that the tax rate set by the government
will not achieve the desired emissions reductions. However, the
importance of this problem is frequently overstated. Because emissions
reductions will take several decades to reach the desired level, the tax rate
can be adjusted over time so that the desired emissions levels are
eventually reached.
Now imagine that you go to the store and are given instructions to
buy twelve cases of soda. In this case, you know exactly how much soda
you will get (twelve cases), but you do not know the cost. In the worst-
case scenario, you might significantly underestimate the cost of a case of
soda and not take enough money – and thereby be unable to get all twelve
cases. This is the situation with a cap and trade. The total number of
permits issued by the government sets the limit on emissions. However,
432

the cost is uncertain. This creates a potential problem for cap-and-trade
systems: Policymakers will issue too few permits (in an effort to bring
emissions down sharply), and the cost of complying will be so high that
significant economic disruption occurs. If this happens, the program might
lose political support and be abandoned. Because of this risk, many cap-
and-trade systems have tended to err on the side of caution and issue far
too many permits. In such a case, the price of the permits goes to zero and
little or no emissions reductions occur.
To address the problem of runaway costs in a cap-and-trade system, a
better approach is to implement an escape valve. Should the cost of
permits rise above a predetermined threshold, the government will sell
more permits at that predetermined price. This would loosen the cap and
increase emissions, but it would also reduce the cost to the economy. If, in
contrast, the government issues too many permits (in an effort to keep
costs down) and the price of permits drops below a predetermined floor,
the government will commit to buying all permits being sold at that price,
thus preventing the price of the permits from falling below that floor.
For political reasons, cap and trade has generally been the preferred
climate policy for the last two decades. The European Union has a cap-
and-trade system operating today, which will be discussed later in this
chapter. In the United States, however, political and economic events
during the first term of the Obama presidency have made cap and trade a
toxic commodity, just like a carbon tax, and there seems to be no chance
that an economy-wide one will be implemented at the federal level. In fact,
at present it appears unlikely that the United States Congress will
implement any comprehensive emissions-reduction policy – in the next
few years, at least.
433

12.2.4 Offsets
Imagine that you own a power plant that emits 100 tons of carbon dioxide
into the atmosphere every year. Imagine that you also own a forest, which
absorbs 100 tons of carbon dioxide from the atmosphere each year. If a
carbon tax is implemented, how much tax should you pay? Do you pay a
tax on emissions (100 tons, from the power plant) or do you get credit for
the carbon dioxide removed from the atmosphere by the forest? If you get
full credit for the forest, then your net emissions are zero because the
emissions from the plant are canceled by the uptake by the forest and you
owe no carbon tax.
Actions that reduce the amount of carbon dioxide in the atmosphere –
you can think of these as “negative emissions” – are often referred to as
offsets. They are called offsets because, if credit is given to these negative
emissions in a climate policy, emitters can use them to offset their
greenhouse-gas emissions, thereby reducing their carbon tax.
From a physics standpoint, offsets should count as negative
emissions. There is no difference to the climate system between not
emitting 1 ton of carbon dioxide and emitting 1 ton while, at the same
time, removing 1 ton via an offsetting mechanism. Offsets make sense
from an efficiency standpoint, too. In much the same way that carbon tax
and cap-and-trade systems are efficient because they encourage the lowest
marginal cost emitters to make the reductions, offsets provide further
flexibility in exactly how emissions are reduced. It may be cheaper, for
example, for a coal-fired power plant to offset emissions by planting trees
than it would be for it to capture the carbon produced in the coal
combustion. Because of this, offsets would be expected to lower the total
cost of reaching any specified emissions target.
434

However, offsets are a much dicier proposition in reality. First, many
offsets are difficult to verify. For example, measuring carbon uptake by a
forest is an extremely complex problem. And what happens if the forest
grows for several years, and then a forest fire burns it to the ground,
releasing the sequestered carbon dioxide back to the atmosphere? How is
this accounted for? Does the forest owner have to refund the payments he
received for the offsets?
Then comes the question of additionality: would the offsetting action
have taken place without the additional value given to the offsetting action
by the carbon emissions regime? To understand what I mean by this,
consider the following example. You own a plot of land, so you go to a
local power plant and offer to plant trees on it if they pay you. They do so,
and in turn they use the carbon absorbed by the growing trees to offset
some of their emissions, which reduces their carbon tax.
The problem arises because we do not know what would have
happened without the payment from the power plant. In order for these
offsets to actually reduce carbon in the atmosphere, the offsetting action
(in this case, planting of trees) should only have occurred because of the
payment from the power plant. If you would have planted those trees
anyway, then the payments from the power plant did not lead to any
reduction in carbon in the atmosphere and should not be used as offsets.
In other words, for an offset to count, we must be sure that the
offsetting actions are in addition to what would have happened anyway
and would not have occurred without the value that they have for climate
change avoidance. Otherwise the offsets achieve no environmental good.
Mitigation programs that include offsets must therefore establish a
mechanism to determine whether an offset satisfies additionality.
For these reasons, offsets turn out to be one of the most controversial
aspects of any mitigation program. In fact, some of the biggest stumbling
435

blocks in the negotiation of the Kyoto Protocol were the proposals to allow
offsets from forests and agricultural lands to satisfy a major part (between
one-fourth and one-half) of the total emissions reductions of each country.
This proposal was pushed by the United States (a country with a lot of
forest and farm lands), but it was steadfastly rejected by Denmark and
Germany.
An aside: The European Trading System
In response to their Kyoto Protocol obligations, the European
Union (EU), an economic and political union of more than two-
dozen European countries, created the EU Emissions Trading
System (ETS) in 2005. The basic structure of the ETS is a cap-and-
trade system, and it covers factories, power plants, and other major
emitters, which make up 40 to 50 percent of the EU’s emissions.
The ETS has gone through several phases since its inception.
During the first phase (2005–2007), almost all permits were given
to businesses free of charge. Because of a lack of knowledge of
historical emissions and worry that an aggressive cap on emissions
might hurt the economy, the number of permits issued exceeded
actual emissions by a large margin, leading to a price for the
permits of zero. Given that emitters reduce emissions until the
marginal cost of the next unit of reduction equals the permit price,
a permit price of zero yielded no incentive for companies to reduce
emissions and effectively no emissions reductions.
The second phase of the ETS covered the Kyoto Protocol’s
commitment period, 2008 to 2012. During this phase, most permits
were still given away free. But in response to the low permit price
during the first phase, the number of permits issued was scaled
436

back. However, 2008 also coincided with the beginning of a
worldwide economic recession, which greatly reduced energy
consumption – and therefore emissions. As a result, even the
reduced number of permits was too high, and the permit price
declined over this period from about $30 per ton to about $6 per
ton – again, the low permit price provided little incentive to reduce
emissions. Emitters were also allowed to use offsets and related
devices to purchase emissions reductions in other countries and use
those in place of reducing their own emissions.
The third phase of the ETS covers 2013 to 2020. Permits will
primarily be issued via auctions, which will help distribute the
permits to those who value them most. There were also a number
of other changes designed to increase the efficiency of the
emissions reductions. Despite these changes, the price of permits
has remained stubbornly low.
Initial problems like those of the ETS are inevitable in any
program as complex as a carbon trading system. And, over time,
they are being worked out. Other cap-and-trade systems are using
the experience of the ETS to help them avoid its mistakes. For
example, California set up a cap-and-trade system in 2013 as part
of a state-level effort to reduce emissions. In order to avoid issuing
too many permits, California regulators studied emissions levels
for several years to determine what the correct number of permits
would be. Permits are issued by a mix of free allocation to certain
industries judged in need of help during the transition to a low-
carbon economy and auction to everyone else. To avoid the price
of permits falling too low, California established a minimum price
of around $11 per ton for the auctioned permits. During the
437

auction, the price of the permits ended up slightly above this
minimum price.
Over time, as more of these trading systems are implemented,
we can expect the world to gain important experience on the best
practices of emissions trading systems.
438

12.3 Information and voluntary
methods
A final way to reduce emissions is simply to give people information. If
people can be convinced that climate change is a serious problem, and then
provided ways to address the problem, they may take some action to
address it without any further prompting by the government.
Information can indeed affect purchasing decisions. For example, car
dealerships in the United States are required by law to put mileage stickers
on cars they are selling. Although not every car buyer is concerned with
mileage, many are, and this information helps them make the socially
beneficial decision of buying a high-mileage car.
In the case of climate change, an example of relevant information is a
greenhouse-gas registry. The requirement to simply report emissions can
provide strong incentives for companies to reduce their emissions.
Companies whose emissions far exceed those of their competitors will be
embarrassed, whereas those with low emission may be viewed as socially
responsible and thereby favored in the marketplace. In both cases, a
registry will give companies incentive to reduce their emissions.
However, this approach has limits. Although informational and
voluntary approaches are quite useful and can be effective at encouraging
people to make some changes to their behavior, these approaches generally
do not compel people to make large or difficult changes. For example,
imagine that your professor asks you to work a few extra problems (that
will not be graded) before the exam. Given the expectation that working
the problems may help you on the exam, and that it is not a large
investment in time, you might choose to do so. Now imagine that your
439

professor asks you to write a twenty-five-page term paper. If the paper
represents a large fraction of your grade, most students will take the
assignment seriously. However, if your professor informs you that she is
not going to grade the term paper, many students will not put much effort
into the paper.
So the government can provide information about climate change and
how to reduce emissions, and it may well cause some people to make some
changes. However, the large changes necessary for us to stabilize the
climate are too big to be motivated simply because we have been told we
ought to make those changes. Thus, informational and voluntary
approaches will likely form part of our response to climate change. They
will not, however, form the fundamental basis for our approach to reduce
emissions.
440

12.4 Chapter summary
The central pillar of most mitigation policies is putting a price on
emissions of greenhouse gases. There are two primary policies to
do this: a carbon tax and a cap-and-trade system.
Under a carbon tax, emitters must pay a tax for each unit of
greenhouse gas emitted. Under a cap-and-trade plan, each emitter
must hold government-issued permits equal to the amount of
greenhouse gases emitted; extra permits can be traded.
Over decades, the tax rate will rise or the number of permits issued
will decrease following a predetermined schedule until the desired
emissions level is reached.
Under these policies, emitters reduce emissions until the marginal
cost (the cost of reducing the next unit) is equal to the carbon tax or
the price of the permit.
These policies are efficient because they shift emissions reductions
to where those reductions can be made most cheaply.
Offsets are processes that remove carbon from the atmosphere –
they can be thought of as negative emissions. Whether these are
allowed to offset real emissions is one of the most contentious parts
of emissions-reduction policy debates. Offsets should satisfy
additionality for them to count. This means that the offsetting
activity would not have occurred without the additional value of
the activity from its impact on emissions.
441

Because of the long lifetime of carbon dioxide, as well as the time
it takes for mitigation policies to reduce emissions, mitigation
efforts we begin today will significantly affect the climate only in
the second half of the twenty-first century.
442

Terms
Additionality
Cap and trade
Carbon tax
Command-and-control regulation
Escape valve
Flexibility
Marginal cost
Offsets
443

Additional reading
There has been a huge amount written about both carbon taxes and cap and
trade. A quick Google search will turn up more books, magazine articles,
and whitepapers than you could ever read. So dive in! What follows here
are a few particular suggestions.
P. Krugman, “Building a Green Economy,” New York Times Magazine,
April 7, 2010. This is a clear and concise summary of the economics of
climate change policy. It pulls together many of the concepts from
Chapters 11 and 12 (download at
www.nytimes.com/2010/04/11/magazine/11Economy-t.html).
W. D. Nordhaus, The Climate Casino (New Haven, CT: Yale University
Press, 2013). This is a much longer, but also more wide-ranging,
discussion of the economics of climate change and the policies to address
it. Chapter 19, in particular, contains a great discussion of the economics
of carbon tax and cap and trade.
See www.andrewdessler.com/chapter12 for additional resources for
this chapter.
444

http://www.andrewdessler.com/chapter12

Problems
1.
a) Explain how a carbon tax works.
b) Explain how a cap-and-trade system works.
c) What is the fundamental difference between these two policies?
d) Given a carbon tax of x dollars (or a permit price of $x), an
emitter will reduce emissions until what criterion is satisfied?
2. Why are emissions reductions achieved by use of a carbon tax or
cap-and-trade system cheaper than those achieved by use of
conventional regulations?
3.
a) What is an offset?
b) What does additionality mean?
4. In a New York Times op-ed piece (December 6, 2009), climate
scientist Jim Hansen makes the following argument: “Consider the
perverse effect cap and trade has on altruistic actions. Say you decide
to buy a small, high-efficiency car. That reduces your emissions, but
not your country’s. Instead it allows somebody else to buy a bigger
S.U.V – because the total emissions are set by the cap.” He argues
that this renders a cap-and-trade system ineffective. Why is this
argument wrong?
5. Imagine a carbon tax is implemented. One day, you decide not to
drive to the grocery store, and you apply for offset credit for the
445

emissions that did not occur because this trip was not taken. Should
you get paid for this? What would you have to prove in order to get
paid?
6. For the following, assume that Plants A and B have the following
marginal costs for reducing emissions:
Number of units
reduced
Marginal costs for
Plant A
Marginal costs for
Plant B
1 3 1
2 5 2
3 7 3
4 9 5
5 11 9
a) The government tells both plants to reduce three units of output.
How much does this “conventional” regulation cost each plant?
What is the total cost?
b) The government implements a carbon tax of $5 per unit. How
much does each plant reduce? What is the total cost?
c) Which approach is cheaper? Why is the cheaper approach
cheaper?
7. The table below shows the marginal costs of two plants, each of
which emits 10 units each year. They both have six permits, meaning
that if they do not trade, they each would have to reduce 4 units.
Assume that they are the only two actors in the market, so the prices
are set by their marginal costs.
446

Number of units
reduced
Marginal costs for
Plant A
Marginal costs for
Plant B
1 1 3
2 2 6
3 3 9
4 4 12
5 5 15
6 6 18
7 7 21
a) How many permits will Plant B buy from Plant A?
b) In what price range will these permits exchange hands?
8. Why will voluntary and informational approaches not lead to deep
reductions in emissions?
9. For a closer-to-home example of a cap-and-trade system, imagine
the following scenario: Your professor gives everyone five points of
extra credit on the final exam. Further, the professor says that you can
sell your extra credit to other students. What would you do? Sell
yours, buy extra, or just hold on to your five points? More generally,
which students will sell their extra credit and which will buy more?
10. When you buy an airline ticket, you can also buy a “carbon
offset” that will cancel out the emissions from the flight. They
typically do not provide much information about the carbon offset.
447

Under what conditions would it be a good thing to buy? Would you
buy one?
11. Another cap-and-trade system is the Regional Greenhouse Gas
Initiative, more commonly referred to by its initials, RGGI (and
pronounced “reggie”). RGGI is made up of several Northeastern
states running from Maine to Maryland. In a paragraph or two,
explain the details of the RGGI and whether it is been a success.
12. One argument made by those who oppose reducing emissions is,
“The energy sources we use are always the cheapest and most
plentiful – which are coal, oil and natural gas. Wind, solar, etc. are
more expensive and therefore bad for the economy.” What is right
and what is wrong about this argument?
1 This proposal was withdrawn in late 2013 when the Obama
Administration released a more comprehensive proposal for regulations
that would also cover existing power plants.
2 For those of you who know calculus, you can think of the marginal
cost as the derivative of the total cost function and the total cost as the
integral of the marginal cost function.
448

13
A brief history of climate
science and politics
◈
In Chapters 11 and 12, we explored our options for addressing climate
change. We can adapt to the change, we can mitigate it by reducing the
emissions of greenhouse gases, or we can geoengineer the climate. In
Chapter 14, I will pull all these together so we can explore how we can
choose among these options. Before we get to that discussion, though, I
describe the context of the policy debate by providing a brief history of
climate change science, policy, and politics.
449

13.1 The beginning of climate science
People have been speculating on the nature of the climate for millennia,
but modern climate science began in earnest two centuries ago, in the early
nineteenth century. In the 1820s, mathematician Joseph Fourier provided
one of the first descriptions of the physics we now know as the greenhouse
effect: A planet’s atmosphere can trap heat and warm the surface of the
planet beyond what it would be if it were a bare, airless rock (we covered
this in Chapter 4). Several decades later, in 1859, physicist John Tyndall
discovered that it was primarily water vapor and carbon dioxide in the
atmosphere that provided the warmth, despite that fact that these two
constituents make up just a small fraction of the atmosphere.
The first recognition that the climate could change occurred in the
1830s, when geologist Louis Agassiz and others identified glacial debris
scattered across Europe. They correctly concluded that northern Europe
must have previously been covered by ice. This was an unanticipated
discovery; prior to that time, everyone had simply assumed that the climate
they experienced was what it had always been and always would be. This
discovery of widespread ice ages showed that climate had changed in the
past, and it certainly suggested that it could change again in the future.
This motivated much of the scientific study of climate over the next
century.
By the end of the nineteenth century, our knowledge of the climate
system was advancing rapidly. In 1896, scientist Svante Arrhenius, a
Nobel Prize winner famous for his studies of chemical reactions, estimated
the climate sensitivity – the warming of the planet from doubling the
amount of carbon dioxide in the atmosphere – and found a value of 5 to
450

6°C. This is a bit higher than modern estimates of climate sensitivity of 1.5
to 4.5°C but still a remarkable achievement given what he knew about the
planet.
Although Arrhenius’ calculations were primarily focused on
explaining the ice ages, he also realized that humans were adding carbon
dioxide to the atmosphere from coal combustion. He estimated, however,
that it would take thousands of years before humans would emit enough
carbon dioxide to significantly warm the climate. He did not appreciate
that fossil fuel use was growing exponentially. As we learned in Chapter
10, the growth of exponentials is so fast that it is easy to underestimate the
long-term trajectory.
The work of Arrhenius really marks the beginning of the theory of
human-induced global warming. But while the bare outlines of modern
climate science were apparent at that time, many fundamental aspects of
climate science were still not well understood. Whereas Arrhenius had
suggested that the carbon dioxide emitted by humans would accumulate in
the atmosphere, many scientists thought that most of the carbon dioxide
emitted by humans would be quickly absorbed by the oceans (as discussed
in Chapter 5, some carbon dioxide is indeed quickly absorbed by the
ocean, but much is not). Furthermore, some scientists suggested that water
so dominated the absorption of infrared radiation by the atmosphere that
adding carbon dioxide would have no effect.
In addition, there was much less concern for environmental issues at
that time. Nature was viewed as dangerous – in fairy tales, children who
wandered into the woods did not come back. Indeed, the twists of weather
and climate were among nature’s cruelest weapons. When a tough winter
or a severe drought could kill you, changing the climate does not seem like
such a big deal.
451

So if the elements of human progress, such as the burning of fossil
fuels, changed the climate, that was okay. Cutting down forest and
replacing it with farmland or hunting predators like wolves to extinction
were considered improvements. Nature was the enemy. Today, of course,
we think differently about nature. We recognize that humans are strong
enough to radically change our environment, and we therefore view the
wilderness as something to be protected and conserved (although
sometimes we do not act that way). “Nature” is somewhere you may go on
vacation, if you can find it and afford to travel there. Figure 13.1
schematically illustrates the power shift between nature and humans over
the past century.
Figure 13.1 Author’s artistic impression of how people viewed their
relationship with nature in the nineteenth century and today.
As an example, consider the fairy tale of Little Red Riding Hood. The
end of that story finds the woodsman killing the wolf with an ax and then
cutting the wolf open and rescuing the grandmother. Unlike today, no one
in the nineteenth century felt bad for the wolf. A modern version of that
fairy tale would end quite differently: Everyone would realize that it was
not the wolf’s fault it ate grandma, so biologists from the U.S. Fish and
452

Wildlife Service would dart and tranquilize the wolf, extract grandma
without injuring the wolf, then radio tag the wolf and release him into a
national park – where he would live happily ever after and (hopefully)
refrain from further consumption of grandmothers.
Temperatures rose during the first few decades of the twentieth
century (see Figure 2.2a), and by the 1930s it was apparent that the planet
was warming. As Time magazine put it in 1939, “gaffers who claim that
winters were harder when they were boys are quite right…weather men
have no doubt that the world at least for the time being is growing
warmer.”1
Around this same time, Guy Stewart Callendar, an English inventor
and engineer, suggested that this warming was caused by human emissions
of carbon dioxide. This is likely the first time that someone suggested that
human-induced climate change was underway. His work built off
Arrhenius’ observation that the burning of fossil fuels would warm the
planet, but he revisited old measurements of atmospheric carbon dioxide
and, unlike Arrhenius, realized that humans were already increasing the
global atmospheric level of carbon dioxide. Like most other people of this
time, however, Callendar was not terribly worried about any detrimental
effects of human modification of the environment.
453

13.2 The emergence of
environmentalism
By the 1950s, our view of the environment was changing as a result of
several factors. One was the invention of nuclear weapons. Nuclear bombs
with a yield of tens of kilotons had been used twice in World War II. By
the 1950s, hydrogen-fusion bombs with yields hundreds of times larger
had been developed. In fact, a single 1950s-era nuclear-armed bomber
could carry bombs with more explosive power than all of the explosives
used in World War II. It dawned on people that we humans now possessed
the power to annihilate ourselves – and people suddenly realized that
humans were now a power in many ways comparable to nature.
Air pollution was also becoming an important issue. Probably the
most famous air-pollution event in history was the killer smog of London
in 1952. In London at that time, most homes were heated with coal, which
dumped sooty smoke into the London air. In early December 1952, a
temperature inversion over London created a stagnant air mass over the
city. As people burned coal, dark soot accumulated above the city. This
dark cloud hung over London, blocking out sunlight, which caused the
temperature to plummet. This caused people to burn more coal for heat,
leading to even more soot in the air. During the height of the event on
Sunday, December 7, the visibility in London was 1 foot. Cattle in the
city’s market were killed and their carcasses discarded rather than sold
because their lungs were black. The particulates harmed people’s health
and killed many of the weak and old. On December 9, the weather
changed and the killer fog was blown away, vanishing as quickly as it had
arrived – but not before several thousand Londoners had died.
454

Air pollution in the United States was bad, too. My father was an
undergraduate at Caltech, and he told me that visibility was so bad in Los
Angeles when he arrived in the fall of 1948 that he was initially unaware
that there were mountains nearby. After he would been there a few weeks,
a rainstorm blew through and enormous mountains a few miles from
campus were suddenly visible. He was, to put it mildly, surprised to see
them. Such persistent air-pollution problems occurred throughout the
world, and this further demonstrated that the human impacts on the
environment were not always for the better.
At the same time, people in many parts of the world were getting
richer. People who are poor tend to worry about where their next meal is
coming from or where there are going to sleep tonight – they are not
terribly concerned with the environment. However, as people become
richer and have disposable income to spend on less essential things,
protecting the environment becomes a higher priority. Once you have
money, you care about your view of the mountains, where you are going to
go camping this weekend, and the extinction of polar bears.
This increasing interest in the environment was bolstered by the
International Geophysical Year, which took place in 1957 and 1958. This
was an international effort that coordinated pole-to-pole observations of
the Earth in order to improve our understanding of the fundamental
geophysical processes that govern the environment. This intensive year of
observations greatly improved our understanding of the Earth – and of the
myriad of ways that humans can alter it. One of the most famous
measurements started during the International Geophysical Year was of
atmospheric carbon dioxide, also known as the Keeling curve (plotted in
Figure 5.6). Within just a few years after commencement, these
measurements showed that atmospheric carbon dioxide was rising as a
455

result of human activities. Here was direct evidence of man’s massive
footprint on the planet.
456

13.2.1 The tobacco strategy
Around this same time, a seemingly unrelated debate was raging in our
society over the heath effects of smoking. By the early 1960s, large
longitudinal studies had proven that smoking cigarettes was associated
with various health risks, such as an increased risk of lung cancer. In 1964,
U.S. Surgeon General Luther Terry issued a report entitled Smoking and
Health,2 which detailed the dire health consequences of smoking. The
evidence supporting their conclusions was enormous – the report
summarized nearly 7,000 scientific articles relating smoking and disease.
The report made newspaper headlines across the country and was the lead
story on television newscasts.
In response, the tobacco companies developed what has become
colloquially known as the tobacco strategy – a concerted effort to cast
doubt on established science in order to advance a particular policy goal.
The goal of the tobacco strategy was not to prove that cigarettes were safe
but rather to create doubt, as described in a tobacco company document
from 1969:
Doubt is our product since it is the best means of competing with the
“body of fact” that exists in the mind of the general public. It is also
the means of establishing a controversy.3
In attempting to generate doubt, the tobacco companies developed a set of
actions to advance their strategy. This included:
Finding a small number of sympathetic scientists who would
convey the message of doubt to the general public. Misrepresent
this to suggest that there is a vigorous debate in the scientific
community.
457

Following this strategy, tobacco companies were able to keep the public
debate over the health impacts of smoking alive for decades after it was
settled in the scientific community. This episode would be a sad historical
footnote if not for the fact that the tobacco playbook has been used again
and again to cast doubt on science suggesting that humans are harming our
environment.
Cherry-pick data and focus on a small number of unexplained or
anomalous details. Ignore the fact that the vast, vast, vast majority
of data solidly supports the consensus view.
Create the impression of controversy simply by asking questions,
even if the answers were known and did not support the tobacco
companies’ case.
Under the guise of fairness, demand equal time from media outlets
to present tobacco companies’ side.
458

13.3 The 1970s and 1980s: Supersonic
airliners, acid rain, and ozone depletion
Environmentalism may have begun in the 1950s, but several events in the
1970s solidified it in the general public’s consciousness. Up first was the
battle over the development of a supersonic airliner. Given the increase in
speed of aircraft from the original Wright Flyer through jet airliners like
the Boeing 707, it was generally believed in the 1960s and 1970s that the
next step for commercial aviation was a supersonic airliner. And the
country that developed the first successful supersonic airliner would garner
enormous national pride as well as economic benefits. Because of this,
Europe, the United States, and the Soviet Union engaged in a three-way
race to develop such an airliner in a competition much like the race to the
moon.
In the early 1970s, however, scientists realized that a fleet of
supersonic airliners might have serious environmental consequences. Jet
engine exhaust includes chemicals that can destroy ozone, and because
supersonic airliners fly at high altitudes for efficiency, these effluents
would be dumped directly into the stratospheric ozone layer. Scientists
began to worry that, given a large enough fleet of these planes, a
significant loss of ozone might result. Because stratospheric ozone blocks
high-energy ultraviolet photons, which harm plants and cause skin cancer
in humans, the loss of this ozone could have serious detrimental effects on
the biosphere.
In the end, the Europeans developed their supersonic airliner, the
Concorde, while the Soviet Union developed their version, the Tu-144.
However, fewer than two dozen Concordes and Tu-144s were ever built.
459

The Concordes flew for three decades, while the Tu-144s flew only a
handful of commercial flights before they were removed from service. The
United States completely abandoned its efforts to build a supersonic
airliner. Although the concern over ozone depletion did play role in
limiting the success of these supersonic planes, it was mainly economics:
supersonic planes are expensive, and consumers would rather have cheap
tickets than fast planes. In fact, today’s modern planes, such as the Boeing
777, actually fly slower than a Boeing 707 – but they trade speed for lower
operating costs, allowing airlines to offer lower prices. As a result, fares in
2005 for flights from New York to Los Angeles – roughly $300 round trip
– were about the same in nominal dollars as they were in the 1940s.
In 1973 and 1974, just as the supersonic transport debate was
subsiding, scientists first theorized that man-made halocarbons might
deplete ozone. The issue of ozone depletion had already entered the public
sphere during the supersonic airliner debate, so policymakers and the
general public were familiar with this risk. At this time, the threat from
halocarbons (frequently referred to as CFCs, which stand for
chlorofluorocarbons) was completely theoretical – it would be more than a
decade before actual observations of ozone depletion were obtained – and
there was no effective replacement for CFCs in many applications.
Nevertheless, in the late 1970s, the United States banned CFCs from being
used in some nonessential applications – such as a propellant in aerosol
spray cans.
In response, CFC manufacturers and industries that used CFCs in
their products joined together to defend the molecule. To do so, they took
the tobacco strategy – attack the science! – and derived a new version
optimized for environmental issues. They focused on three main claims: It
is not happening; if it is happening, we are not to blame; and, if we are to
blame, then fixing the problem will be too expensive. As with the tobacco
460

debate, the goal of those defending CFCs was not to prove any of these
points, but to simply generate enough uncertainty and doubt to stall any
policy to regulate CFCs.
At this same time, another environmental problem was emerging:
acid rain. Many power plants, particularly those that burn coal, emit large
amounts of sulfur dioxide and nitrogen oxides to the atmosphere. Once in
the atmosphere, these molecules can be absorbed by cloud droplets and
raindrops, and once dissolved in water they react to form sulfuric acid and
nitric acid. This is analogous to the way carbon dioxide dissolves into
water to form carbonic acid (discussed in Chapter 5). The difference is that
carbonic acid is what is known as a weak acid, whereas nitric and sulfuric
acids are strong acids. This means that dissolving nitrogen oxides or sulfur
dioxide into water makes the resulting liquid much more acidic than
dissolving an equal number of molecules of carbon dioxide. When this
acid rain falls to the ground, the types of potential damage it can do are
numerous: bleaching of nutrients from soils, acidification of lakes and
rivers, damage to wildlife and plants, damage to human-built structures,
and so on.
This entire theory of acid rain is scientifically quite simple, and
research done over the 1970s and 1980s definitively connected emissions
from power plants to the acidic precipitation. In response to this research,
the first broad international agreement covering acid rain, The Convention
on Long Range Transboundary Air Pollution, was signed in Geneva by
thirty-four member countries of the U.N. Economic Commission for
Europe on November 16, 1979. The next year, the Council of the
European Communities enacted a directive reducing sulfur dioxide
emissions.
The Reagan Administration, however, was resistant to enacting
regulations on emissions of acid-rain precursors. To support their
461

reticence, the administration created a series of reports that argued that
there was too much uncertainty to take action:
“The state of the science…probably will not yield a scientifically
complete assessment of acid deposition in the next few years,” says
the report prepared for Congress and the Reagan Administration. “To
date, the state of the science will not allow assertive
recommendations. Trends are weak and evasive. Data are spotty. One
of the most basic uncertainties is the extent of damage caused by acid
deposition and its rate of change.” The report says air pollutants,
some of which cause acid rain, are the prime suspects in investigation
of dying forests in the northeastern United States but “the association
between damage and the occurrence of those pollutants is not well
defined.” It says some lakes have been acidified and fish have died
but “the rate, character and full extent of these changes remain major
scientific unknowns.” And it says acidity can accelerate corrosion of
buildings but “information about effects on materials from acid
deposition needs to be better defined.”4
The United States took no action on acid rain during the Reagan
Administration. However, in the early 1990s, the George H. W. Bush
Administration enacted regulations to reduce sulfur emissions through a
cap-and-trade system – just like the cap-and-trade systems discussed in
Chapter 12. These regulations greatly reduced the emissions of sulfur at a
price far below expectation. And, as expected, this greatly decreased the
occurrence of acid rain. The success of cap and trade in helping solve the
acid rain problem is one of the main reasons that policymakers looked
hopefully at that mechanism for addressing climate change.
It is worth reiterating that the cap-and-trade solution in the United
States to the acid rain problem emerged from the Republican George H.
462

W. Bush Administration. At the time, it was viewed as a conservative-
friendly method of solving the problem because it let the market determine
how emissions reductions would be allocated. For this same reason,
environmentalists were suspicious or outright opposed to cap and trade. It
was unclear if it would actually reduce acid rain, and many
environmentalists viewed the program as allowing emitters to pay to
pollute. If pollution was wrong, they reasoned, you should not be able to
pay to do it. Over time, and with the incredible success of the program,
environmentalists came around to viewing cap and trade as a particularly
effective way to solve environmental problems. This is why they started
thinking of solving the climate change problem using this same policy
instrument. At the same time, Republicans have developed a steadfast
opposition to these cap-and-trade schemes.
The ozone problem remained an active scientific and political issue
throughout the early and mid-1980s. The original theories from the 1970s
suggested that ozone depletion would be a slow process, taking half a
century or longer for significant depletion of ozone to occur, and that it
would primarily occur at mid-latitudes and high altitudes. But when
scientists obtained the first evidence that ozone was actually being
depleted as a result of CFCs, they found it was occurring much more
rapidly and in a different place than predicted. They observed extremely
rapid loss of ozone over Antarctica, where every spring, roughly 90
percent of the ozone in the lower stratosphere was being destroyed in a
month or so (it built back up during the rest of the year so that it was
available to be destroyed again the following year). This annual loss of
ozone became known as the ozone hole.
Within a few years, newly discovered chemical reactions that relied
on CFCs combined with the unique meteorology of the polar regions were
identified as the cause of this rapid polar ozone depletion. This confirmed
463

the role of humans and suggested that this problem might be more serious
than had previously been recognized. In response to this threat, the world
adopted the Montreal Protocol in 1988, an international agreement
committing the world to phasing out CFCs.
An important aspect of the Montreal Protocol was that the phase-out
of CFCs happened in two stages. Industrialized countries phased out CFCs
first, followed ten years later by developing countries. There are several
reasons for this. Industrialized countries are richer than the developing
countries, so they have more resources to apply to phasing out CFCs.
Moreover, by having the rich countries go first, economies of scale and
technical advances would be expected to drive down the cost for
developing countries of phasing out CFCs. There were also ethical
considerations. The CFCs in the atmosphere – which were causing the
ozone depletion – had mainly been released to the atmosphere by activities
in the industrialized countries. Developing countries had contributed little
to the problem. Thus, it was agreed that industrialized countries should
take the first step to clean up the problem.
During and after the negotiation of the Montreal Protocol, the science
of ozone depletion was advancing rapidly, and evidence continued to
accrue about the dangers CFCs posed to the ozone layer. But even as the
science became more certain, so-called ozone skeptics stepped up their
attacks on the science of ozone depletion. A good example is this 1989
quote from National Review:
The current situation can fairly be summarized as follows: The CFC-
ozone theory is quite incomplete and cannot as yet be relied on to
make predictions. The natural sources of stratospheric ozone have not
yet been delineated, theoretically or experimentally. The Antarctic
ozone hole is ephemeral; it comes and goes, and seems to be
464

controlled by climatic factors outside of human control rather than by
CFCs.
The reported decline in global ozone may be an artifact of the
analysis. Even if real, its cause may be related to the declining
strength of solar activity rather than to CFCs. The steady increase in
malignant melanoma has been going for at least 50 years and has
nothing to do with ozone or CFCs. And the incidence of ordinary skin
tumors has been greatly overstated.…And substituting for CFCs is no
simple matter. A New York Times report of March 7, 1989 talks
about the disadvantages of the CFC substitutes. They may be toxic,
flammable, and corrosive; and they certainly won’t work as well.
They’ll reduce the energy efficiency of appliances such as
refrigerators, and they’ll deteriorate, requiring frequent replacement.
Nor is this all; about $135 billion of equipment use CFCs in the U.S.
alone, and much of this equipment will have to be replaced or
modified to work well with the CFC substitutes. Eventually that will
involve 100 million home refrigerators, the air-conditioners in 90
million cars, and the central air-conditioning plants in 100,000 large
buildings. Good luck! The total costs haven’t really been added up
yet.5
If this argument sounds familiar, it is because it is the tobacco strategy.
And in retrospect, we now know that all of these arguments are wrong.
Two decades of research have concretely verified the link between CFCs
and stratospheric ozone depletion. What is more, the costs of replacing
CFCs with ozone-safe alternatives turned out to be so small that, when
CFCs were completely phased out in the mid-1990s, virtually no one
noticed.
In addition to the scientific uncertainty arguments, the ozone skeptics
also provided a context for why so many scientists, politicians, and
465

activists were convinced by such shoddy science:
It’s not difficult to understand some of the motivations. For scientists:
recognition for keeping dusty records or running complicated
computer models that are rather dull; more grants for research; press
conferences; and newspaper stories. Also the feeling that maybe they
are saving the world for future generations. For bureaucrats the
rewards are obvious. For diplomats there are negotiations, initialing
of agreements, and – the ultimate – ratification of treaties. It doesn’t
really much matter what the treaty is about, but it helps if it supports
“good things.” For all involved there is of course travel to pleasant
places, good hotels, international fellowship, and more. It’s certainly
not a zero sum game.
I have left environmental activists to the last. There are well-
intentioned individuals who are sincerely concerned about what they
perceive as a critical danger to the health of future generations. Many
of the professionals share the same incentives as government
bureaucrats: status, salaries, perks and power. And then there are
probably those with hidden agenda of their own – not just to “save the
environment” but to change our economic system. The telltale signs
are the attack on free enterprise, the corporation, the profit motive, the
new technologies. Some are socialists, some are Luddites.
Most of these “compulsive utopians” have a great desire to regulate
– on as large a scale as possible. To them global regulation is the
“holy grail.” That’s what makes the CFC-ozone issue so attractive to
them.
Thus, an alternative narrative is created: The science the public is hearing
in the mainstream media is wrong, and the reason wrong science is being
conveyed is that scientists and advocates are corrupt, biased, or stupid. No
466

evidence is provided to support these charges, but it is not really necessary.
The point here is to cast doubt, not to prove the accuracy of this narrative.
The Earth’s temperature remained relatively constant between the
1940s and 1970s (Figure 2.2). Despite this, research since the 1950s had
emphasized the risk of global warming due to increasing abundances of
atmospheric greenhouse gases. At the same time, however, the abundance
of aerosols from human activities was also rising (e.g., from the burning of
high-sulfur coal). As discussed in Chapter 6, aerosols can cool the planet,
which offsets some of the warming from increased greenhouse gases.
Some scientists suggested that humans were in fact adding enough
aerosols to the atmosphere to overpower greenhouse gases and that the
dominant influence of man was a net cooling of the climate.
A legitimate scientific debate ensued over which effect would
dominate, and by the end of the 1970s the debate had been settled in favor
of those predicting that global warming would be the dominant human
influence. In recent years, some have misrepresented this mere existence
of a debate to suggest that the scientific community in the 1970s was
predicting global cooling. This is incorrect – there was never any
widespread consensus among scientists that aerosol-induced cooling was
the dominant influence of humans.6
The 1970s ended with the publication of an influential report7 from
the U.S. National Academy of Sciences that reviewed the science and
came to this conclusion: “If carbon dioxide continues to increase, the study
group finds no reason to doubt that climate changes will result and no
reason to believe that these changes will be negligible. The conclusions of
prior studies have been generally reaffirmed.” More quantitatively, they
concluded that a doubling of carbon dioxide would result in a warming of
1.5 to 4.5°C, which is the same as today’s estimates. More research in the
early 1980s fleshed out and confirmed the general view that humans were
467

in the process of modifying the climate. However, the general public and
most politicians were not yet focused on the issue.
468

13.4 The year everything changed: 1988
1988 was the year when climate change went from being a mostly
academic problem to a political one. The United States was blisteringly hot
that summer, with much of the country suffering drought conditions and
temperature records smashed on a seemingly daily basis. A small number
of U.S. congressional leaders were interested in the problem of climate
change, and they felt the time was right to hold a congressional hearing on
it. In a stroke of political genius, they held the hearing in August, which is
the hottest part of the Washington summer. Committee staffers opened the
windows of the hearing room the night before, so the room itself was also
extremely hot. Videos of the hearing show people sweating and mopping
their brow, effectively reinforcing the message about climate change.
At that hearing, NASA climate scientist James Hansen declared that
he was 99 percent confident that the world really was getting warmer and
that there was a high degree of probability that it was due to human
activities. Coming on the heels of the publicity over the ozone hole, this
created a media firestorm, and it put the issue of climate change onto the
political radar. In the next few months, the United Nations passed a
resolution urging the “Protection of global climate for present and future
generations of mankind.” Time magazine, instead of naming a “Person of
the Year” for 1988, named “Endangered Earth” the “Planet of the Year.”
This was also the year that the Intergovernmental Panel on Climate
Change, or IPCC, was formed (this organization was discussed in Chapter
1). During the negotiation of the Montreal Protocol, the World
Meteorological Organization and the U.N. Environmental Program had put
out a series of reports that described the science of stratospheric ozone
469

depletion. These assessments were incredibly successful at establishing the
bedrock scientific principles upon which the Montreal Protocol was based.
The climate policy community, seeing the success of these ozone
assessments, concluded that similar assessments about climate science
would help facilitate those policy debates, and this was the job the IPCC
was created to do.
As momentum for enacting regulations to reduce emissions began to
grow, so did the pushback from those opposed to regulations. Given that
the energy is a several trillion dollar per year business, it should come as
little surprise that many people and institutions strenuously opposed
regulations that would cost them some of these trillions.
They were joined by those philosophically opposed to environmental
regulations. This group was mainly motivated by the fundamental belief
that regulations on greenhouse gas emissions were an unacceptable
infringement on freedom. This is summed nicely up by Vaclav Klaus,
President of the Czech Republic (and one of the very few leaders of any
country to doubt the mainstream view of the science of climate change):
“The largest threat to freedom, democracy, the market economy, and
prosperity at the end of the twentieth and at the beginning of the twenty-
first century is no longer socialism. It is, instead, the ambitious, arrogant,
unscrupulous ideology of environmentalism.”8
With the tobacco strategy in mind, those opposed to regulations on
greenhouse gases focused on attacking the science. To do this, they
recruited a small group of contrarian scientists to make the public
argument. Many of these so-called climate skeptics were veterans of
previous battles – tobacco, acid rain, ozone – and they had deep experience
casting doubt. The skeptics’ views of science are heterogeneous; for
example, some skeptics dispute that the Earth is warming, while others
accept that the Earth is warming but dispute that humans are responsible.
470

Some even dismiss the fundamental physics of the greenhouse effect. But
they typically share a disdain for any science that might lead to increased
government regulation.
In 1990, the IPCC put out its first assessment on the science of
climate change. In it, the IPCC concluded that “the size of this [observed]
warming is broadly consistent with predictions of climate models, but it is
also of the same magnitude as natural climate variability. Thus the
observed increase could be largely due to this natural variability.” This
relatively weak statement about the role of humans in climate change
reflected legitimate uncertainties in climate science at that time. Because
of these uncertainties, a definitive attribution of the warming to
greenhouse gases was not possible.
471

13.5 The Framework Convention on
Climate Change: The First Climate
Treaty
Despite this, many world leaders felt that action had to be taken on climate
change. The result was the Earth Summit in Rio de Janeiro in 1992, from
which emerged the treaty known as the Framework Convention on Climate
Change, frequently referred to simply as the “framework convention” or
by its initials, FCCC. The FCCC enjoys near-universal membership, with
192 countries having ratified it, including the United States, China, and all
other big emitters. The principles enshrined in the FCCC remain the major
building blocks on which negotiations of treaties to reduce emissions have
been built.
The most contentious debate over climate change policies involves
mitigation. In that regard, the stated goal of the FCCC is “to achieve
stabilization of greenhouse gas concentrations in the atmosphere at a low
enough level to prevent dangerous anthropogenic interference with the
climate system.” This is fine as far as it goes, and it receives widespread
agreement at this level of abstraction. In practice, though, the meaning of
this statement hinges on the definition of the word dangerous. There is no
scientific definition of what dangerous climate change is because this is a
value judgment – climate change that one person may perceive as
dangerous may not be perceived that way by someone else.
In order to bring fairness or equity to any climate change agreement,
the FCCC also enshrines the concept of common but differentiated
responsibilities. This means that all countries must participate in solving
the climate change problem, but not necessarily the same way. For
472

example, we might expect rich, industrialized countries to begin cutting
their emissions first, with developing nations cutting their emissions later.
The reasons for this are similar to the reasons that, in the Montreal
Protocol, industrialized countries phased out CFCs first, followed ten years
later by developing countries. The industrialized countries of the world are
far richer than the developing countries, so they have more resources to
apply to reducing emissions. Moreover, by having the rich countries go
first, economies of scale and technological advancement would bring
down the cost of reducing emissions so that, when developing countries
did begin reducing their emissions, the cost to them would be less.
There are also moral considerations. The 2 billion or so poorest
people in the world currently live hard lives of crushing poverty. One of
the ways to raise these people out of poverty is through economic growth –
increasing their consumption of goods and services. This requires energy,
so anything that makes consuming energy harder or more expensive for
the poorest will also make it harder to lift these people out of poverty.
Common but differentiated responsibility is a way of saying that solutions
to climate change should not work at cross-purposes to efforts to reduce
poverty.
There is also the question of historical responsibility. Most of the
increase in carbon dioxide in the atmosphere over the past 250 years is due
to emissions from the rich and industrialized countries. In fact, the world’s
rich countries are rich because of the energy they consumed – and the
emissions that resulted. Thus, it makes sense for them to have a greater
responsibility for taking the first steps toward cleaning up the problem. It
is also clear, however, that developing countries must eventually
contribute. China is now the largest emitter of carbon dioxide, and several
other developing countries are either presently major emitters or on track
to be. Reducing global emissions significantly over the coming century
473

will be impossible without these developing countries eventually making
deep emissions reductions.
The FCCC also included what is referred to as the precautionary
principle: “Where there are threats of serious or irreversible damage, lack
of full scientific certainty should not be used as a reason for postponing
such measures.” This is a crucially important and widely misunderstood
statement about the implications of scientific uncertainty in policy
deliberations. It does NOT say that, if the risks are high enough, we must
take action, regardless of scientific uncertainty. Rather, it says that, if the
impacts of a risk are sufficiently serious, scientific uncertainty should not
be used as an excuse to do nothing. There may be many other reasons to
do nothing (e.g., economic, moral), but science should not be one.
The FCCC was intended to be a starting point for more specific and
binding measures to be negotiated later. Consequently, in contrast to its
ambitious principles and objectives, the treaty’s concrete measures were
weak. Under the FCCC, nations committed to reporting their current and
projected emissions and supporting climate research. Parties also accepted
a general obligation to adopt, and report on, measures to limit emissions.
What these measures had to be, or had to achieve, was not specified. Only
for the industrialized countries did this general obligation also include the
specific aim of returning emissions to 1990 levels by 2000. This target was
nonbinding, meaning that it was an aspirational target and there were no
sanctions for missing the target.
474

13.6 The Kyoto Protocol
By the middle of the 1990s, it was clear that no country would achieve the
emissions-reduction target set in the FCCC and that a treaty with
mandatory reductions would be required. Around that same time, the IPCC
released its second assessment of the science of climate change, in which it
concluded that “the balance of evidence suggests a discernible human
influence on the climate.” This quote reflected the fact that the science of
climate change had significantly advanced since the IPCC’s first
assessment and there was now much more evidence linking the observed
warming to human activities. But significant uncertainties remained.
In response to these developments, the Kyoto Protocol was negotiated
in December 1997. Unlike the FCCC’s nonbinding emissions reductions,
the Kyoto Protocol required emissions from participating industrialized
countries, averaged over a commitment period running from 2008 through
2012, to be approximately 5 percent below their 1990 emissions level.
Developing countries, however, had no emissions reduction requirements.
The Protocol incorporated several provisions to allow flexibility in
how nations met their emission limits. As I discussed in Chapter 12,
flexibility mechanisms allow emissions reductions to be shifted to where
they can be made most cheaply, thereby reducing the overall cost of
attaining a particular emissions target. Flexibility mechanisms included
targets being defined for total emissions of a basket of carbon dioxide and
five other greenhouse gases (methane, nitrous oxide, hydrofluorocarbons,
perfluorocarbons, and sulfur hexafluoride). Countries could meet their
target by reducing emissions of any of these gases, not just carbon dioxide.
475

Emissions-reduction obligations could also be exchanged between
nations through various mechanisms. Under one mechanism, known as the
Clean Development Mechanism, industrialized countries could invest in
emissions reduction projects where it is cheapest globally – typically in
developing countries – and count those reductions toward their own goal.
For example, France can invest in a wind farm in China and count the
reduction in emissions toward France’s emissions target. Like offsets, such
projects must satisfy additionality: it must be demonstrated that the wind
farm in China would not have been built without the financial support of
France through this program.
The Protocol also included provisions for nations to meet some of
their obligation through offsets. The Protocol included credit for
reforestation, but some countries, such as the United States, wanted credit
for other carbon-capturing activities, such as agriculture. This turned into a
significant conflict at a negotiating session in November 2000 in The
Hague. A proposed compromise was almost reached, but it was ultimately
rejected at the last minute by the French and German environment
ministers, who judged that the proposed offsets weakened the Kyoto
commitments too much.
476

13.7 The George W. Bush years:
2001–2008
Shortly after the breakdown in negotiations at the meeting in The Hague
and just a few months after taking office in 2001, the Bush Administration
announced it was withdrawing the United States from the Kyoto Protocol
process. The reasons included too much scientific uncertainty about
climate change and potential harm to the U.S. economy. Although it later
retreated from claiming that its withdrawal was based on scientific
uncertainty, the Bush Administration continued to hold that the Protocol
was unacceptable because of the high costs to the U.S. economy.
A particular problem cited by the Bush Administration was the
absence of emission limits for developing countries. Although “common
but differentiated responsibilities” is enshrined in the FCCC – which
George H. W. Bush signed – and had also been incorporated into the
Montreal Protocol, the George W. Bush Administration painted this as
unfair to the United States. While it is true that China was already a major
economic competitor to the United States in many areas, China was also a
much poorer country than the United States and had far fewer resources to
devote to reducing emissions.
Also in 2001, the IPCC released its third assessment report on the
science of climate change. The report came to this conclusion: “There is
new and stronger evidence that most of the warming observed over the last
50 years is likely attributable to human activities.” Compared with the
previous reports, this one made a much more definitive statement about the
role of humans in the recent warming. However, there was still
477

uncertainty, as reflected by the use of the word likely, which denotes a two
out of three chance.
In February 2002, President Bush outlined his alternative approach to
the issue. While consistently saying that climate change was a problem
that needed to be addressed, the Bush Administration steadfastly avoided
talking about reducing greenhouse-gas emissions. Instead, the emphasis
was on reducing the greenhouse-gas intensity – the T term of the IPAT
relation, which concerns emissions per dollar of GDP. Their stated goal
was to reduce greenhouse-gas intensity by 18 percent by 2012. This was a
weak goal because greenhouse-gas intensity has historically declined at 1
to 2 percent/year without any policies. Thus, the Bush goal would likely be
met with little or no effort.
The Bush policies also increased funding for climate change science
and for specific technologies to reduce emissions, implemented tax
incentives for renewable energy and high-efficiency vehicles, and started
several programs to encourage voluntary emission cuts by businesses.
While the Bush Administration made no serious effort to reduce
emissions, the rest of the industrialized world continued pushing for the
Kyoto Protocol. To enter into force – and so become binding on those who
ratified – the Protocol required ratifications by fifty-five countries,
including nations contributing at least 55 percent of 1990 industrialized-
country emissions. This threshold meant that, after the withdrawal of the
United States, the treaty could enter into force only if all other major
industrialized countries joined. The fate of the Protocol remained uncertain
until November 2004 when, after several years of uncertainty about its
intentions, Russia submitted its ratification, allowing the Protocol to enter
into force on February 16, 2005.
But the long delay awaiting the required ratifications meant that the
Protocol entered into force only three years before the start of the five-year
478

commitment period (2008–2012). Some nations, such as the European
Union, took aggressive action by enacting a large-scale cap-and-trade
program, the European Trading System (ETS) (this was discussed in detail
in the previous chapter). Other countries made little effort to achieve what
would be a large deflection of emissions over very few years.
At the same time, in the absence of any federal efforts to reduce
emissions in the United States, efforts trickled down to the state and local
levels. For example, Connecticut, Delaware, Maine, Maryland,
Massachusetts, New Hampshire, New Jersey, New York, Rhode Island,
and Vermont banded together to form the Regional Greenhouse Gas
Initiative, more commonly referred to by its initials, RGGI (and
pronounced “reggie”).9 The RGGI is a regional cap-and-trade program that
covers emissions from just one economic sector – electric power plants.
The Western Climate Initiative was formed by a group of western states
and Canadian provinces to also develop a regional cap-and-trade system.
In addition, many individual U.S. states and cities began efforts to reduce
emissions.
In 2007, the IPCC released its fourth assessment on the science of
climate change and came to this conclusion: “Most of the observed
increase in globally averaged temperatures since the mid-twentieth century
is very likely due to the observed increase in anthropogenic greenhouse
gas concentrations.” This continued the trend toward stronger statements
implicating humans in the warming – the words very likely here denote 90
percent confidence.
With the end of the Kyoto Protocol in 2012 in sight, representatives
of the world’s governments met in Bali in December 2007 to begin
negotiations for a new climate treaty that would build on what the Kyoto
Protocol had accomplished, to be ready before a meeting in Copenhagen in
2009. Importantly, it was agreed in Bali that this new agreement would,
479

unlike the Kyoto Protocol, include emissions reductions by both
industrialized and developing countries. Subsequent negotiations quickly
split over the relative efforts required of these two groups.
In 2008, as George W. Bush’s presidency was reaching its end, the
campaign to succeed him heated up. Senator Barack Obama was the
Democratic nominee, and he was opposed by the Republican nominee,
Senator John McCain. Both candidates accepted the reality of climate
change and the need to do something about it – in fact, Senator McCain
had tried several times over the previous decade to get an emissions
reduction bill through the U.S. Senate. The disagreements between the
candidates on climate policy were quite minor – e.g., arguments over how
much we should rely on nuclear power to reduce emissions, how deep the
cuts should be in 2050.
480

13.8 The Obama years: 2009–today
Barack Obama became president of the United States in January 2009.
Shortly thereafter, in December 2009, an international meeting in
Copenhagen took place to negotiate the follow-on to the Kyoto Protocol.
While hopes were raised by renewed U.S. engagement under the Obama
Administration, the Copenhagen meeting was marked by continuing
disputes between developing and industrialized countries over sharing the
burden of action. Developing nations wanted the industrialized world to
make sharp, near-term (e.g., by 2020) reductions in emissions, whereas the
industrialized world wanted the developing nations to agree to quantitative
emissions reductions.
On the final day of the conference, President Obama and a handful of
key developing country leaders negotiated an agreement known as the
Copenhagen Accord. The Accord included these major points:
Global temperatures should not rise more than 2°C beyond
preindustrial temperatures.
Deep cuts in emissions will be necessary. Recognizing equity
issues, these deep cuts may be delayed in developing and poor
countries.
The world’s rich industrialized countries each agreed to set their
own target for emissions in 2020. Since the adoption of the Accord,
the European Union, for example, agreed to a reduction of 20 to 30
percent below 1990 emissions, while the United States agreed to a
reduction of 17 percent below 2005 emissions.
481

Around that same time, Obama and his congressional allies began
advancing major bills on health care and climate change through Congress.
In response to this, as well as general opposition to President Obama’s
policies, the Tea Party, a libertarian wing of the Republican Party
dedicated to reducing the role of government in our lives, became a major
force in U.S. politics. Because policies to reduce emissions require some
government intervention in the energy market – usually by pricing carbon
emissions either through a carbon tax or a cap-and-trade system – the Tea
Party rabidly opposes climate legislation.
The rise of the Tea Party put immense pressure on Republican
politicians to reject climate change science and any legislation to reduce
emissions. Before 2009, a number of prominent Republicans openly
acknowledged the risk of climate change and supported policies to reduce
greenhouse-gas emissions. This included leaders in the party such as John
McCain (2008 presidential nominee), Mitt Romney (2012 presidential
nominee), and Newt Gingrich (former speaker of the house). After 2009,
however, each of these politicians adopted a skeptical position on the
science of climate change and opposed legislation to reduce emissions.
The world’s developing countries agreed to take on mitigation
efforts but did not accept specific emissions targets. China, for
example, agreed to reduce greenhouse gas intensity (the T term in
the IPAT relation) by 40 to 45 percent below 2005 levels by 2020.
However, given the rapid economic growth of China, this still
corresponds to increased emissions.
Flexibility should be incorporated into policies in order to achieve
emissions reductions at the lowest cost.
Adaptation must necessarily be part of our response. Industrialized
countries agree to provide resources to help poorer countries adapt.
482

Republican politicians who did not adopt this viewpoint quickly found
themselves out of a job.
In 2010, Tea Party-affiliated candidates, virtually all of whom reject
the science of climate change or the seriousness of the problem, were
elected in numbers high enough to fundamentally change the composition
of Congress. This ended any opportunity to get comprehensive climate
legislation through the U.S. Congress.
Conservative pushback on climate policy was also occurring in other
countries. In Canada, the conservative government of Prime Minister
Stephen Harper withdrew from the Kyoto Protocol in 2011. The
government did this because of well-worn criticisms of the Protocol (e.g.,
it does not include developing countries) and because Canada has immense
oil reserves in the form of tar sands, which will likely be worth
significantly less if the world agrees to stringent emissions reductions. In
2013, a conservative government in Australia began rolling back a price on
carbon emissions that had been implemented several years earlier by the
previous government.
By the end of 2012, the end of the Kyoto Protocol’s commitment
period, most industrialized countries had not achieved their Kyoto Protocol
targets (2008–2012 emissions about 5 percent below 1990 levels). Those
that did, mostly in Central and Eastern Europe, relied on the fact that the
base year was 1990, prior to the collapse of the Soviet Union. After the
dissolution of the Soviet Union in 1991, much of the inefficient industry
there was shut down, leading to a huge decrease in emissions; despite
subsequent growth, emissions there had not yet climbed back to 1990
levels by the Kyoto Protocol’s commitment period.
In 2012, Barack Obama was reelected president, and he made climate
change a key issue in his second term. Given the composition of Congress,
though, getting any climate legislation through it was clearly impossible.
483

The Obama Administration instead turned to executive orders – policies
the president can implement without congressional approval – and existing
authority granted to the administration under the U.S. Clean Air Act to
address climate change.
In 2014, the EPA announced regulations for fossil-fuel-powered
electricity generating units. The required emissions reductions vary by
state but will lead to an average reduction of about 30 percent from these
plants. Owing to the high carbon intensity of coal, these new regulations
will make it quite difficult to operate conventional coal-fired power plants.
These regulations will be challenged in court, but, if enacted, they
represent the first legitimate effort by the United States to reduce
emissions.
Around the same time, the IPCC released its Fifth Assessment
Report. It concluded that, “It is extremely likely that human influence has
been the dominant cause of the observed warming since the mid-20th
century.” This continues the trend towards strengthening the attribution
statement – it now uses the words extremely likely, which denotes a 95
percent chance (compared to very likely (90%) in the Fourth Assessment
and likely (66%) in the Third Assessment).
484

13.9 The breakthrough: U.S.-China
bilateral agreement
In November 2014, a blockbuster climate deal was announced. The United
States and China, the two biggest emitters, together responsible for about
one-third of emissions, agreed to limit their emissions. The United States
agreed to emit 26 to 28 percent less carbon dioxide in 2025 than it did in
2005. This was perhaps not such an amazing commitment: in response to
the Copenhagen Accord, the United States had already committed to
reducing emissions by 2020 – but this agreement doubled the pace of
emissions reductions.
More impressively, China agreed that its emissions would peak
before 2030. This was a true shift in policy, as China had previously only
talked about greenhouse gas intensity reductions, not emissions reductions.
In order to do this, they committed to produce 20 percent of their energy in
2030 from renewable power sources. This requires them to build about 1
GW of renewable power every week for the next fifteen years.
While the deal only covers the United States and China, I believe that
future generations will look upon this as a turning point in negotiations
over international climate policy. Including the E.U., which already has
aggressive emissions reductions targets in place, this deal means more than
half of the world economy, which emits more than half of the carbon
dioxide, has agreed to significant deflection of their emissions trajectories.
Most importantly, China is the de facto leader of the developing
world, and China’s agreement to limit emissions in the near term will put
enormous pressure on other developing countries to do the same. It will
put even more pressure on industrialized countries that refuse to reduce
485

emissions. Over the years, the excuse that China had steadfastly refused to
reduce emissions had become one of the most important arguments for
those countries – and now that excuse is gone. Countries like Canada and
Australia will find it harder now to do nothing about climate change.
486

13.10 Chapter summary
Scientists have been studying climate change for nearly 200 years,
and in that time a successful theory of climate has emerged. This
theory is described in Chapters 1–7 of this book.
The first prediction of human-induced climate change was made by
Svante Arrhenius, who recognized in the late nineteenth century
that human combustion of fossil fuels might warm the climate. In
the late 1930s, Guy Stewart Callendar made the first claim that
human-induced global warming had arrived.
In the 1950s, people realized that humans possessed the power to
greatly modify our environment – and not to our benefit. And the
economic growth and increases in wealth occurring at that time
meant the environment had more value to people, and people had
more money to spend to enjoy it.
In the 1970s and 1980s, the debates over ozone depletion and acid
rain were a preview for the debate over climate. Those opposed to
action on these problems refined the strategy of the tobacco
companies: Cast doubt on the science.
The first climate treaty was the Framework Convention on Climate
Change or FCCC. This treaty enshrined three important principles:
1) “common but differentiated responsibilities,” 2) the
precautionary principle, and 3) an agreement that the world should
limit greenhouse-gas emissions in order to prevent “dangerous”
climate change.
487

The 1997 Kyoto Protocol included binding reductions of emissions
for industrialized countries – these countries had to reduce
emissions from 2008 through 2012 by roughly 5 percent below
1990 emissions. There were no restrictions placed on developing
countries.
The 2009 Copenhagen Accord included the agreement that the
world’s goal should be to avoid 2°C of warming above
preindustrial temperatures. In addition, industrialized countries
agreed to set their own emissions reduction targets for the year
2020.
In late 2014, the United States and China mutually agreed to limit
their emissions. This may be viewed in the future as a turning point
in the efforts to get an international agreement to limit emissions.
488

Terms
Chlorofluorocarbons, or CFCs
Climate skeptics
Common but differentiated responsibilities
Copenhagen Accord
Equity
Framework Convention on Climate Change, or FCCC
Killer smog of London
Kyoto Protocol
Montreal Protocol
Ozone hole
Precautionary principle
Tobacco strategy
489

Additional reading
S. R. Weart, The Discovery of Global Warming, 2nd ed. (Cambridge, MA:
Harvard University Press, 2008). This is an accessible, well-written
historical timeline of primary developments in the science of climate
change, from the 1800s through the formation of the modern consensus
about the predominantly human cause of recent climate change as
expressed in the recent IPCC reports (accessible online at
www.aip.org/history/climate/index.htm).
N. Oreskes and E. M. Conway, Merchants of Doubt: How a Handful of
Scientists Obscured the Truth on Issues from Tobacco Smoke to Global
Warming (London: Bloomsbury Press, 2010). As I stated in Chapter 1, this
important book explains how deception is used to misguide the public on
various matters, from the risks of smoking to ozone depletion to the reality
of global warming.
See www.andrewdessler.com/chapter13 for additional resources for
this chapter.
490

http://www.aip.org/history/climate/index.htm

http://www.andrewdessler.com/chapter13

Problems
1. Your roommates have a party when you are out of town.
a) When you return, the apartment is a mess and they ask you to
help clean it up. Do you help them?
b) Under what conditions might you offer to help? What might
they offer you to get you to help them?
c) How is this situation analogous to the debate between
developing and industrialized countries over mitigation efforts?
2. Who was the first person to discover that climate could change?
Who was the first person to predict that human emissions of carbon
dioxide might warm the climate? Who first claimed that human-
induced climate change was occurring?
3. Do an Internet search and find some Web sites skeptical of
mainstream climate science. List three of the claims they make about
the science of climate change. Given what we have covered in the
first twelve chapters of this book, are these arguments convincing?
4. What are the four important components of the Framework
Convention on Climate Change?
5. What difference is there in how we view the environment today
versus how people who lived in the nineteenth century viewed it?
What are the factors that caused the change?
6. Explain the precautionary principle. Can you think of an example
in your life when you have applied the concept (or explicitly not
applied it)?
491

7.
a) Explain the concept of equity as it was described in this chapter.
b) How was it implemented in the Montreal Protocol?
c) Give an example of how it might be implemented in a climate
agreement.
8. Under the Copenhagen Accord, China agreed to reduce its
greenhouse gas intensity by 45 percent in 2020 (relative to a base year
of 2005).
a) What annual growth rate gives you a 45 percent decrease over
fifteen years?
b) If affluence grows at 7 percent/year and population grows at 1
percent/year, what would be the expected change in total emissions
over the entire time period?
9. How does the bilateral U.S.-China climate agreement incorporate
the concept of equity?
1 Quoted in Weart (2008).
2 profiles.nlm.nih.gov/NN/B/B/M/Q/
3 A large number of tobacco company documents can be viewed on the
Legacy Tobacco Documents Library (see legacy.library.ucsf.edu/). This
particular document can be found at
legacy.library.ucsf.edu/tid/wjh13f00/pdf.
4 “Acid Rain Facts Called Sketchy,” The Globe and Mail (Canada),
June 12, 1984.
492

http://profiles.nlm.nih.gov/NN/B/B/M/Q/

http://legacy.library.ucsf.edu/

http://legacy.library.ucsf.edu/tid/wjh13f00/pdf

5 S. F. Singer, “My Adventures in the Ozone Layer,” National Review,
June 1989.
6 For a good review of the history of “global cooling” and how it is
misrepresented in today’s debate, see Peterson et al., “The Myth of the
1970s Global Cooling Scientific Consensus,” Bulletin of the American
Meteorological Society 89 (2008): 1325–1337.
7 Ad Hoc Study Group on Carbon Dioxide and Climate, Carbon
Dioxide and Climate: A Scientific Assessment (Washington, DC:
Climate Research Board, National Research Council, 1979).
8 V. Klaus, Blue Planet in Green Shackles (Washington, DC:
Competitive Enterprise Institute, 2007).
9 In May 2011, New Jersey announced its intention to leave the
program.
493

14
Putting it together: A long-
term policy to address climate
change
◈
We have now reached the final chapter on our trip through the problem of
modern climate change. In the previous thirteen chapters, we explored the
fundamental physics that leads us to confidently conclude that humans are
now changing the climate and that continuing to add greenhouse gases to
the atmosphere could bring serious changes to our climate over the next
century and beyond. We are not certain how bad this climate change will
be, but the upper end of the range (global and annual average warming of
4°C or more over the twenty-first century) includes warming large enough
for the experts to consider its impacts to be potentially catastrophic. Even
the lower end of the range, about 1°C, is more warming than occurred
during the twentieth century and will be challenging for the world’s
poorest as well as our most vulnerable ecosystems. We have also explored
a number of possible responses to this risk, including mitigation,
adaptation, and geoengineering. We have even touched briefly on the
political debate over climate change.
494

In this chapter, I will discuss the elements of an effective response to
climate change. I will also show the step-by-step logic that underlies the
most commonly suggested policies for addressing climate change. Our
choice of climate and energy policy must reflect the science as well as the
economic trade-offs and moral judgments about the alternatives.
495

14.1 What makes climate change such a
difficult problem?
I hope that, by this point, you recognize that the physics of the climate
problem is actually pretty simple and we can have high confidence that
humans are altering the climate. While there is uncertainty in the science,
it is mainly about how bad the impacts of climate change will be, not
whether humans are altering the climate or whether negative impacts will
occur.
Despite the scientific certainty, it has been difficult for the world to
get together and address this threat. To see what makes this such a hard
problem, let us first consider a different, well-known problem: terrorism.
Probably the most famous terrorist attack in history took place on
September 11, 2001, when terrorists hijacked and crashed four U.S.
airliners, flying three of them into buildings and killing about 3,000
people. In that attack, there was a clear perpetrator (the hijackers and their
facilitators) and a clear victim (the people killed and survivors who are
terrorized); the cause and effect were clear (the hijackers took control of
the planes and crashed them, killing the victims and terrorizing survivors);
the effect was immediate (the impacts occurred immediately after the
hijackings); and there was a clear intent to do something bad (the hijackers
intended to harm people and society).
The response to this attack was immediate and strong. The United
States invaded Afghanistan and toppled the government, who had aided
the 9/11 terrorists. The attack also provided a key motivation for the 2003
U.S. invasion of Iraq. Given the strong and aggressive response to the
terrorist attacks of 9/11, why does another existential threat, climate
496

change, get so little traction in U.S. policy discussions? The answer is that
climate is a more diffuse problem than terrorism. First, we cannot pick out
a single perpetrator. The people responsible are just about everyone who
has lived since the industrial revolution, including you and me. In addition,
no single person emits enough greenhouse gases to be an issue. If you
were the only emitter, there would be no problem. It is only because there
are billions of people on the Earth that emissions are causing the climate to
change.
Second, the victims of climate change are also dispersed in time.
While we are already altering the climate, the most severe impacts of
climate change will be felt toward the end of the twenty-first century and
beyond, when the climate is several degrees Celsius warmer than today.
This means that many of the people who will be harmed by climate change
have not even been born. There is also a disconnect between cause and
effect in space: many of the impacted will live in very poor countries,
while most emissions come from richer countries. The upshot of this is
that many people alive today do not feel personally threatened by climate
change the way they do by terrorism. This is probably the most important
reason that climate is not a higher policy priority than it is.
Third, the link between cause and effect is not as straightforward as it
was for the 9/11 attacks. While we can reasonably attribute rises in global
average temperatures to human activities, it is more difficult to attribute
individual events that really affect people, such as individual heat waves,
droughts, floods, and severe storms to climate change due to human
activities.
Finally, those who cause climate change (i.e., everyone) do not intend
to cause it. You do not get into your car with the intent of causing climate
change, and you would still drive your car even if it did not produce
greenhouse gas emissions (in fact, most people would prefer it if their car
497

did not cause climate change). Rather, emissions and the resulting climate
change are a side effect of economic activities that are otherwise virtuous –
this means that restrictions on emissions carry the possibility of harming
our economy, so one must balance climate policies against restrictions in
economic activity. Overall, these differences make the climate problem
harder to deal with than more immediate threats.
498

14.2 Decisions under uncertainty:
Should we reduce emissions?
Despite the difficult philosophical aspects of the problem discussed in the
last section, we nevertheless face a choice with climate change: to act or
not to act. That decision will be based on answers to a few key questions.
Some of these are scientific (How much warming will we experience?
What will the impacts be?), some are economic (How expensive will it be
to respond?), and some are (at least partially) moral (How bad will the
impacts be? Is geoengineering an acceptable response?). Ultimately, we
don’t have precise answers to any of these questions. Such uncertainty is
not unusual in important policy decisions. Most important decisions
(Should we invade Iraq? Should we cut taxes? Should we implement
universal health care?) are made with incomplete information. So how can
we make a decision in the face of this uncertainty? One way to think about
this problem is to consider the following two arguments:
Both statements argue that we must err on the side of caution in order to
avoid a bad outcome. However, the bad outcome is different in these two
arguments. In the first argument, the bad outcome is severe damage from
climate change, whereas in the second it is severe economic damage from
responding to climate change.
Because the worst-case scenario of climate change is so serious, we
must take action now to reduce emissions, even though we don’t
know exactly how bad climate change will be.
Because of the high cost of reducing emissions, we must be certain
that climate change is serious before we take action.
499

So which of these arguments is correct? We can gain some insight
into how to think about this by looking at some familiar examples of
decisions in the face of uncertainty. First, consider a criminal trial. To
convict someone of a crime, a jury must be convinced that the defendant is
guilty beyond a reasonable doubt. The reason for this standard is that we,
as a society, have decided that it is better to acquit a guilty person than
convict an innocent one.
Put another way, there are two errors a jury can make. They can
convict an innocent person or they can acquit a guilty one. These errors are
not equally bad – we judge that it is worse to convict an innocent person.
So the standard of conviction (“guilty beyond a reasonable doubt”) is set to
minimize the possibility of making the worse mistake. In doing so, we
increase the chance that we make the other mistake, acquitting a guilty
person.
Another example occurs in deliberations concerning national defense.
For example, former Vice President Dick Cheney famously said, “If
there’s a one-percent chance that Pakistani scientists are helping al Qaeda
build or develop a nuclear weapon, we have to treat it as a certainty in
terms of our response.”1 As in our jury example, there are two errors here
that we could make. We could respond as if al Qaeda had a nuclear
weapon, but it turns out they do not have a nuclear weapon. Or we do not
respond, and it turns out they do have one. In most deliberations about
national defense, being unprepared for a threat is judged to be a worse
error than to respond to a threat that never materializes. That is the
fundamental judgment that Cheney is making here.
So how do we think about climate change? Must we be certain
beyond a reasonable doubt that climate change is a serious threat to
mankind before taking action to reduce emissions? Or should we take
action to reduce emissions even if there is just a 1 percent chance that it is
500

a serious threat? This question boils down to your judgment of which error
is worse: reducing emissions unnecessarily because climate change turns
out to be a minor threat or not reducing emissions and climate change
turns out to be a serious threat.
Suppose that climate change turns out to be a minor problem. In that
case, an aggressive mitigation program would impose costs as we rebuild
our energy infrastructure from fossil fuels to renewables and other climate-
safe energy sources. How bad would this be? Switching from fossil fuels
to climate-safe energy has advantages completely unrelated to climate,
such as reductions in air pollution. Moreover, because costs of
transitioning to climate-safe energy would be spread over the next several
decades, at least some of the cost can be avoided by scaling back future
efforts once we learn they are unnecessary. Furthermore, fossil fuels will
be exhausted in the next century or so. Thus, switching away from fossil
fuels is inevitable and these costs are going to be paid eventually. The
bottom line is that it is hard to imagine a person living in Year 2100 being
upset that we switched from fossil fuels to climate-safe energy sources.
Now suppose that climate change turns out to follow the worst-case
scenario. If we do nothing to reduce emissions, we doom the planet to
much warmer temperatures for the next millennium and beyond. This
could impose catastrophic costs on our society (both economic and moral).
In this case, it is easy to imagine that a person living in Year 2100 would
be furious that we did nothing to address a problem that we clearly saw
coming. In the end, it is difficult to make the argument that taking too
much action on climate change is a worse error than taking too little. This
would suggest a standard closer to Cheney’s, that the risk of climate
change justifies action to reduce emissions, even in the face of significant
uncertainty.
501

Another factor that enters into decisions under uncertainty is
irreversibility. If an action you take is irreversible, you have to be more
certain that it is the right action than for a decision that is easily reversible.
That is why, for example, inmates on death row in the United States are
allowed many appeals to their death sentence – executing someone is as
irreversible an action as there is, so you want to be as sure as possible that
you have executing the right person. Just putting someone in jail, in
contrast, is reversible – if you realize later that you have made a mistake,
you can simply release that person.
Reducing emissions is a reversible decision. If we decide later that
climate change is not that serious, we can always change our policies and
increase our consumption of fossil fuels. And the investments we make in
alternative energy sources, such as solar and wind, are reversible over a
few decades.
But the converse is not true. Once you emit carbon dioxide, there is
no practical way to remove it from the atmosphere. Instead, you have
committed the planet to millennia of higher temperatures or
geoengineering. Many of the impacts of climate change, such as extinction
of species, sea-level rise, and loss of the Greenland or Antarctic ice sheets,
are irreversible on time scales that we care about. Thus, the irreversibility
of emitted carbon dioxide and its associated climate change tends to favor
taking action to reduce emissions. In the end, most people who have
seriously looked at the problem, including almost every world
government, have concluded that action on climate change, in particular
the reduction in emissions, is justified given the risks.
502

14.3 Picking a long-term goal
Any emissions-reduction effort requires a long-term goal. The deeper the
cuts in emissions, the less climate change we will eventually experience –
but the more expensive those cuts will be. This is the trade-off, and we
want to pick a target that avoids the worst climate change but at a cost that
is manageable and does not interfere with other policy goals, such as
poverty reduction.
The Framework Convention on Climate Change says that we should
strive to avoid “dangerous” climate change. But what is dangerous? It is
not a scientific term, so science provides input, but cannot settle the issue.
Rather, it is a value judgment that takes science into account as well as our
views on topics such as risk, poverty, environmental stewardship,
government regulation, and many other contentious topics. There are
various ways to determine a long-term goal, and I discuss two of them in
this section.
503

14.3.1 Cost versus benefits
In Chapter 12, we described how a company would respond if a price were
put on emissions through a carbon tax or cap-and-trade system. As
described there, companies compare costs and benefits to choose the
emissions reduction that maximizes their net benefit (benefits minus cost).
This occurs when the cost of reducing one more unit (the marginal cost)
equals the cost of emitting that unit (either through a tax on that unit or the
cost of a permit to emit that unit).
Societies can make emissions reductions decisions in an analogous
way: by comparing the cost of various levels of emissions reduction to the
benefits obtained by making those reductions. Imagine, for example, that
our economy can reduce emissions of carbon by 1 ton for $5, but we get
$50 of benefits by avoiding the climate impacts of that ton. From a societal
standpoint, that is a no-brainer – we should certainly not emit that ton.
Now imagine that the next ton of emissions can be avoided for $6, and
avoiding emission of this ton delivers $49 of benefits. Again, it is clear
that we should pay to not emit that ton either. As we cut emissions deeper,
the cost of eliminating each subsequent ton (the marginal cost) rises, while
the benefits from avoiding that ton (the marginal benefit) decline. This is
the law of diminishing returns, which we explored in Chapter 12.
The reduction that gives us the largest net benefit would be our
preferred goal, and this occurs when emissions are reduced until the cost
of reducing one more ton equals the benefit from avoiding that ton. Even
though this may be conceptually straightforward, the actual calculation is
not. For example, we know how much renewable energy costs today, so
we can estimate the cost of replacing fossil fuel energy. But we also know
that putting a price on carbon will spur development of new technologies
504

by providing a financial incentive that does not exist in today’s economy.
That this will happen is not in doubt – the question is how fast the
innovation will take place. If we assume that innovation responds rapidly
to putting a price on emissions, then the cost of reducing emissions will be
much lower than if we do not make that assumption. Depending on what
assumptions are made for this and other uncertainties, estimates of the
costs of reducing emissions cover a wide range; some analyses conclude
that it will be quite cheap whereas others conclude that it will be ruinously
expensive.
Estimating the benefits from avoiding climate change is even more
difficult. First, we do not at present have the ability to predict changes in
temperature and precipitation at the regional scales required for detailed
estimates of impacts. Second, converting estimated changes in climate into
a dollar figure can be difficult and arbitrary. For goods and services that
are traded in markets (e.g., food, lumber, recreational skiing), calculating
the economic loss due to climate change is relatively straightforward. For
things that are not traded in markets, however, estimates of the cost of
climate change are far more arbitrary. Take, for example, the extinction of
polar bears. Polar bears do not contribute much to the global economy, so
their extinction would likely have a negligible financial cost. But many
people nevertheless value polar bears and would view their loss as a
significant harm. Economic analyses can attempt to quantify the value of
polar bears by using methodology that is neither uniform nor entirely
satisfactory, or analyses can ignore their loss, which implicitly assigns it a
value of zero – which is also unsatisfactory. Given the many problems in
estimating the cost of climate change, it should come as no surprise that
there is also a wide range of cost estimates.
Another problem in estimating the costs of climate change comes
from the timing of climate impacts. If 1 ton of carbon is emitted into the
505

atmosphere today, it will warm the planet for centuries to come and will
cause impacts over that entire time. But the cost of avoiding those impacts
– by not emitting that ton – must be paid today. To compare costs that are
occurring at different times, we therefore need to convert the value of the
climate impacts over the next few hundred years to its value to us today
(i.e., its present value). This involves discounting, which we explored in
detail in Section 10.4.
Although discounting is conceptually straightforward, the big
uncertainty is what discount rate to use. Most analyses in the climate
change policy debate use discount rates between 0 percent and 4 percent.
The larger the discount rate, the lower the present value of future costs –
and we will consequently be willing to pay less to avoid those impacts.
For example, with a discount rate of 0 percent, $1 trillion of climate
change damage in 100 years has a present value of $1 trillion. We should
therefore be willing to pay up to $1 trillion dollars today to avoid it. In
contrast, if we select a discount rate of 4 percent, then $1 trillion of climate
change damage in 100 years has a present value of $20 billion – meaning
we would only be willing to pay $20 billion to avoid those impacts.
This is even more problematic for impacts occurring far in the future.
As we explored in Chapter 8, our emissions this century will commit the
planet to warming and climate impacts over the coming millennium and
beyond. The present value of $1 trillion of climate impacts in 1,000 years
at a 0 percent discount rate is $1 trillion; at 4 percent, it is about 0.001
cent. The staggering difference between these present values has
significant implications for how much emissions reduction is optimal.
Together, uncertainties in estimates of the costs of reducing
emissions, costs of climate impacts, and the discount rate lead to a wide
range of estimates of how deeply to cut emissions in order to maximize net
benefits. Some estimates are for deep, immediate cuts, while others call for
506

more gradual ramping down of emissions over decades. Because of this,
economics is limited in its ability to prescribe a quantitative response to
climate change.
Nonetheless, there are some things that all economic analyses agree
on, and on those points we can have high confidence. There is widespread
agreement that reductions in emissions make sense and that carbon pricing
should be the centerpiece of the policy. In addition, there is agreement that
the price on emissions should rise with time, eventually becoming very
high. The net result is that emissions can be allowed to grow in the near
term (for perhaps a decade or so) before emissions must actually begin to
decline. The exact timing of the emissions peak is determined by how far
we decide emissions must be cut – the more climate change we want to
avoid, the nearer in time the cuts must begin. It is also important that we
have near-universal participation by all countries in any climate regime, in
order to make the necessary reductions at the lowest cost.
Economic analysis suffers from another problem: It looks at
aggregate costs and benefits but not their distribution. It does not take into
account that many of the hardest-hit regions are also the poorest regions of
the world – regions that have contributed little to climate change and have
the least resources available to address climate. Many would view that
outcome as fundamentally unfair, a factor not considered in economics. So
although economics tells us what the most efficient solution is, it tells us
nothing about whether the solution is fair or just.
A final problem with economic analysis comes from the problem of
catastrophe, including worst-case outcomes such as abrupt climate
changes, mass starvation, or even human extinction. Economic analyses
struggle to assign a monetary value to a small and hard-to-quantify risk of
truly terrible outcomes like these. As a result, most economic analyses
ignore the worst-case scenario and focus on the most likely outcomes. This
507

is a significant shortcoming: For many people, the mere possibility of true
catastrophe in the next few centuries provides sufficient motivation for us
to address climate change now.
508

14.3.2 Target: 2°C
Given all of the problems with economic analyses, expecting them to
quantitatively determine the optimal amount of mitigation is not realistic.
A simpler way is to choose some threshold in the climate system that we
should avoid. The threshold should give us a good chance to avoid serious
climate impacts, but it should be relaxed enough that it is politically and
economically acceptable. Over the past few years, a consensus has grown
around limiting warming to 2°C above preindustrial temperatures. This is
what was adopted, for example, at the FCCC’s Copenhagen meeting in
2009. The scientific argument in favor of this threshold is that modern
human society, with megacities and large-scale industrial food production,
developed during a period of relatively small climate variations – less than
2°C. This stable climate provided the conditions (e.g., good weather for
agriculture, robust freshwater supplies) under which human society
flourished.
Ultimately, though, this 2°C is arbitrary. There is no scientific
analysis that proves that 2°C is the most appropriate target, nor any reason
to think that warming slightly below this threshold is much better than
warming slightly above. And we could just as easily have chosen a
completely different metric, such as a limit in rate of warming (how many
degrees per century we were willing to accept) or a limit in atmospheric
carbon dioxide. Some advocates, for example, argue that we should limit
atmospheric carbon dioxide to 350 or 450 ppm. However, as of 2014,
consensus has arisen around the 2°C target, and it has most of the
momentum in policy discussions.
509

14.4 How do we get there?
510

14.4.1 The physics of a 2°C limit
This is a challenging target – we have already experienced approximately
0.8°C of warming above the preindustrial level, and we are committed to
another few tenths of a degree Celsius of warming from emissions that
have already occurred. This leaves us a bit less than 1°C of warming that it
is still physically possible to avoid.
Avoiding all of that 1°C is impossible because that would require
turning off most fossil-fuel energy immediately. That would be too
expensive – the best we can hope for is to phase fossil fuel power out over
the next few decades, during which time atmospheric carbon dioxide will
continue to build up. But how fast do we need to turn off fossil fuels?
A useful way to think about this problem is in terms of a carbon
budget. Because carbon dioxide remains in the atmosphere for so long
(discussed in Chapter 5), the peak warming we will eventually experience
is proportional to cumulative emissions of carbon dioxide. Climate models
suggest that the climate will warm about 2°C for every trillion tons of
carbon emitted (note: 1 trillion tons of carbon = 1,000 billion tons = 1,000
GtC). Humans have already emitted about 500 GtC since the industrial
revolution, so we can stay under 2°C warming if we limit future emissions
to 500 GtC. This is a tight budget: While it took us 250 years to plow
through the first half of our carbon budget, given the exponential growth
of the rate of emissions, we are on schedule to plow through the second
half of our carbon budget in the next few decades – unless we take action
to reduce emissions.
Other greenhouse gases matter, of course, but carbon dioxide is the
most important because it remains in the atmosphere for so long (discussed
in Chapter 5). For a gas like methane, with a ten-year lifetime, once you
reduce emissions, the atmospheric perturbation due to humans, and the
511

associated radiative forcing and warming, will fall to zero within a few
decades.
As with all budget problems like this, we have three options: we can
start cutting emissions now and reduce our emissions (relatively)
gradually, or we can continue to do nothing, let emissions grow each year,
and then make drastic emissions cuts later. Or we can do nothing now and
do nothing later and simply accept large changes in our climate.
A few numbers will help clarify our choices. If we choose to do
nothing, warming over the twenty-first century could be 4°C (this is the
prediction of the RCP8.5 scenario, see Figure 8.5), corresponding to about
5°C warming above preindustrial. That is about as much warming as we
have had since the end of the last ice age. And warming is expected to
continue well beyond 2100 under this scenario. Given such significant
warming, the risks of severe climate impacts would be considerable.
If we decide that we want to keep global warming below 2°C – and
we decide we want to start immediately – what level of effort would be
required? In this case, we can keep cumulative emissions below 500 GtC if
we reduce emissions by 1.5 percent/year. So how do we do that? I noted in
Chapter 8 how emissions can be thought of as the product of population,
affluence, and technology (Equation 8.1: I = P × A × T; if you do not
remember this, now is a good time to review Section 8.1). We expect the
world’s population to continue to increase in the future, and we also expect
that people will become richer. More quantitatively, we can expect the
product of population times affluence (P × A) to increase at 2 to 4
percent/year. This means that the technology term (i.e., the greenhouse gas
intensity) must decline by 3.5 to 5.5 percent/year in order for overall
emissions to decline by 1.5 percent/year.
Again thinking back to Chapter 8, greenhouse gas intensity is equal to
the energy intensity times the carbon intensity (equation 8.2). We expect
512

that energy intensity, a measure of how efficiently an economy uses
energy, to continue to decrease by about 1 percent/year. Thus, the bulk of
our emissions reduction comes from the carbon intensity term, which must
decrease at 2.5 to 4.5 percent/year.
The carbon intensity term is a measure of the technology used to
generate electricity. To have it decline by 2.5 to 4.5 percent/year will
require construction of about 1 gigawatt (1 GW = 1012 W) of carbon-free
energy every day for the next century. For scale, 1 GW is about the power
produced by a coal-fired power plant, a large hydroelectric dam, a few
hundred large wind turbines, or a square of solar panels that is a few
kilometers on each side.
It is easy to conclude that this task is too big and that we cannot
possibly build this much carbon-free energy. But the power of human
industrial society is also immense – it is, after all, big enough to shift our
climate. As an example, consider the following fact: human population
increased by about 1 billion in the last thirteen years (corresponding to a
population increase of 200,000 people/day) and most of the population
increase occurred in urban areas. This means that humans have been
building the equivalent of a city of one million people every five days for
the last decade. And we will continue this rate of urban construction for
the next several decades, at least. Given that, switching to carbon-free
energy does not seem so daunting, and it is clear that we could build
enough carbon-free energy if we chose to do so.
If we wait to start reducing emissions until, say, 2030, we have to
reduce greenhouse gas intensity faster – about 6 to 11 percent/year – to
stay within our 2°C carbon budget. If we wait until 2040 to start reducing,
then we have to reduce the greenhouse gas intensity even faster – by 13 to
50 percent/year. These all correspond to much higher rates of building
carbon-free energy sources and higher overall costs. Given how difficult it
513

has been for the world to agree to do anything at all about climate, I am
skeptical that the world would ever agree to such high rates of emissions
reductions. In such a situation, it seems likely to me that we would turn to
geoengineering to buy time to let a more modest mitigation program take
effect.
514

14.4.2 How to get there from here
As described in the previous section, we know what needs to happen from
a scientific viewpoint in order to stabilize the climate at 2°C above the
preindustrial level – we need to cut emissions of greenhouse gases,
particularly carbon dioxide, and we need to start as soon as possible. It is
also clear that all countries need to participate in this endeavour. If one or
a few big countries choose not to participate, then participating countries
will have to make larger and faster cuts in emissions. This is more
expensive for the overall economy, and it is particularly expensive to those
making the steep reductions. This leads to more economic pain for those
participants, potentially eroding political support in those countries.
But getting all nations to agree to reduce greenhouse gas emissions is
difficult because, in the international arena, no one is in charge. There is
no world government with the authority to implement and enforce policy
or to compel governments to participate. Rather, international policy is
made by negotiation among national representatives. This process is
weaker, more cumbersome, and slower than national policy-making, but it
is all that is available to respond to global problems such as climate
change.
The goal of international climate negotiations is for each country to
commit to some level of effort to address climate change. These national
commitments could be performance targets, such as limits on national
emissions, leaving it up to the individual governments how to achieve
them. Alternatively, governments could commit to enact national policies,
such as emission taxes or other forms of regulation. Or they could submit
to international processes to motivate and enable national actions, such as
515

reporting and exchange of information, assessment of national policies, or
reviews of their progress.
Thus far, national emissions targets, in which countries agree to
reduce emissions by some specified amount in some specified time period,
have been the most common form of international mitigation commitment,
used in both the Framework Convention on Climate Change and the Kyoto
Protocol, as well as several other major environmental treaties. National
targets have the advantages of being simple, clear, and familiar. Moreover,
by defining clear responsibilities but leaving the means of implementation
up to national governments, targets let governments be held accountable
for their commitments with minimal intrusion on their sovereignty.
But national targets also have serious disadvantages. Because most
emissions come not from governments but from citizens and businesses, a
national target has no concrete effect until implemented in domestic
policy. But the outcomes of domestic policies are uncertain at the outset,
so governments cannot know in advance how difficult or costly it will be
to meet an emissions target, or even whether a particular target is
achievable. This uncertainty about targets’ attainability introduces a sharp
tension into international target negotiations. If governments face only
minor consequences for missing a target, they have little incentive for
serious efforts to meet it. But if the consequences are severe, governments
will likely only agree to targets they are highly confident of meeting –
weak targets, or ones with wide loopholes. Moreover, while clear,
demanding targets can be good motivators, they do this best when nations
are near the boundary between meeting and not meeting them. Incentives
to exceed a target you already expect to meet, or to narrow the gap when
you are clearly falling short, are much weaker.
But the biggest problem with national targets is assigning burdens and
costs among nations. The most intense disputes in this category are
516

between developing countries and the rich, industrialized countries.
Developing countries argue that the burden for near-term cuts should fall
on the world’s rich countries, and they advance several reasons in support
of this. The first is historical responsibility. Most of the increase in carbon
dioxide in the atmosphere over the past 250 years is due to emissions from
the industrialized countries. In fact, the world’s richest countries are rich
because they have consumed a lot of energy. Thus, it makes sense for them
to have a greater responsibility for taking the first steps toward addressing
the problem.
Second, the industrialized countries are wealthier and have more
resources available to make the investments required to reduce emissions.
By having the rich countries reduce emissions first, development of new
technologies would be expected to lower the cost for the developing
countries of their eventual emissions reductions.
Third, climate change policies cannot work against poverty reduction.
Two billion people today live in intense poverty. Lifting them out of that
state requires energy, so anything that makes consuming energy harder or
more expensive for the poorest will also make it harder to reduce poverty.
This is a strong moral argument that the poorest countries in the world
should not be asked to reduce emissions. As they become richer, they can
contribute to emissions reductions.
The argument that the rich world should move first has received
mixed reviews in the rich world. In 1997, in response to the negotiation of
the Kyoto Protocol, the U.S. Senate passed a resolution with a vote of
95–0 stating that the United States should not mandate new commitments
to limit or reduce greenhouse gas emissions unless the agreement also
includes commitments to limit or reduce greenhouse gas emissions by
developing countries.
517

Over the following seventeen years, this issue was never really
settled. Developing countries refused to agree to reduce emissions,
although they agreed that they would have to do so eventually. The
European Union moved ahead on its own aggressive emissions reduction
targets, while Canada and Australia essentially abandoned any formal
targets. The United States was somewhere in between. During Obama’s
second term, various executive orders (which do not require congressional
approval) were enacted that are expected to reduce emissions, although
they fell short of the kind of comprehensive energy policy that would
significantly reduce U.S. emissions. Nevertheless, U.S. emissions have
been declining since the mid-2000s. This is mainly due to three factors: 1)
the great recession, which slowed the growth of affluence, 2) a decrease in
energy intensity due to the long-term increases in energy efficiency of the
economy, and 3) the decrease in carbon intensity due to a shift away from
coal following the rise of new drilling technology (i.e, fracking), which
flooded the energy market with cheap shale gas.
As I write this in late 2014, important events have signalled a possible
breakthrough in international negotiations. As discussed in the previous
chapter, the United States and China recently announced a bilateral
agreement to rein in emissions: The United States intends to reduce its
emissions by 26 to 28 percent below its 2005 level by 2025. China intends
to achieve the peaking of carbon dioxide emissions around 2030 and to
make efforts to peak early. China also intends to increase the share of non-
fossil fuels in primary energy consumption to around 20 percent by 2030,
which will require the construction of about 1 GW of renewable power
every week between now and then. The agreement between the de facto
leaders of the industrialized and developing worlds to limit emissions has
the possibility of breaking the rich versus poor divide that has stymied
negotiations over the past twenty years.
518

14.4.3 What policies should look like
The national targets that result from international negotiations do nothing
until implemented in domestic policy. So what would a domestic policy
look like? The first and most important action every nation can take is to
put a price on emissions of carbon dioxide and other greenhouse gases
(why this is necessary was covered in Chapter 11). Economist William
Nordhaus put it this way:
Whether someone is serious about tackling the global warming
problem can be readily gauged by listening to what he or she says
about the carbon price. Suppose you hear a public figure who speaks
eloquently of the perils of global warming and proposes that the
nation should move urgently to slow climate change. Suppose that
person proposes regulating the fuel efficiency of cars, or requiring
high-efficiency light bulbs, or subsidizing ethanol, or providing
research support for solar power – but nowhere does the proposal
raise the price of carbon. You should conclude that the proposal is not
really serious and does not recognize the central economic message
about how to slow climate change. To a first approximation, raising
the price of carbon is a necessary and sufficient step for tackling
global warming. The rest is at best rhetoric and may actually be
harmful in inducing economic inefficiencies.2
Second, although a price on carbon is crucial, there are some economic
sectors where great progress can be made rapidly but where expected
progress with just a price on carbon will be slow. For these sectors,
efficiency standards and other incentives can be implemented to encourage
careful energy use. One example is fuel mileage standards for automobiles.
519

In addition, restrictions on activities that are particularly unfriendly to the
climate, such as the burning of coal, might also be considered.
Third, countries should fund the research and development of new
technologies. Although it may be possible to solve the climate problem
with existing technology, it is also clear that new and improved
technologies can ease the transition as well as reduce the cost. An example
of transformative technology would be a breakthrough in battery energy
storage. This would reduce the impact of the intermittency of renewable
energy – e.g., solar energy generated during the day could be stored and
then fed into the grid at night. Another example would be an “artificial
leaf” that takes sunlight and uses it to split water into hydrogen and
oxygen; the hydrogen could then be burned to produce energy.
Fourth, prepare to adapt to climate change. Regardless of what
actions we take now, the globe will continue to warm for decades. And to
the extent that this warming cannot be avoided, we must adapt to it. Thus,
adaptation must necessarily be a part of our response to climate change.
Anticipatory adaptation is cheapest, so we should begin to incorporate the
reality of climate change into any plans for the future. This is particularly
true for investments in long-lived infrastructure. When we build a road,
airport, power plant, or the like, we must make sure that it is resilient to
any reasonable changes in the climate. In addition, not every country has
the resources to adapt, so mechanisms of providing international aid may
have to be implemented.
Fifth, there is also the unfortunate possibility that the world will not
get its act together to reduce emissions. If the world does nothing to reduce
emissions in the next few decades, then geoengineering may be our only
way to avoid truly disastrous warming. We should prepare today for this
by researching the geoengineering options in order to determine which are
most likely to work and what the negative side effects might be.
520

Sixth, we must also realize that whatever policy we adopt now will
probably not be exactly the right long-term policy. Because of this, both
international agreements and the domestic policies that flow from them
must be reviewed and amended as new information about the science of
climate change and new technological developments arise.
In the end, we do not confidently know how hard it will be to reduce
our emissions and make the transition to a fossil-fuel-free future. In this
way, the climate change challenge is not unique; we almost never know in
advance how much it will cost to comply with environmental regulations.
Before regulations to reduce ozone depletion were passed in the 1980s, for
example, some advocates predicted that the regulations would cause an
economic apocalypse, with people in the developed world having to get rid
of their air conditioners and millions in the developing world dying
because of a lack of food refrigeration.
It turned out that innovation in response to the threat of regulation led
to the development of substitutes for the ozone-depleting chemicals so
cheap that, when the new chemicals replaced the older ones, virtually no
one noticed. The hope with climate change is that, once a price is put on
carbon emissions, innovation by the private sector will produce
breakthrough technologies that allow us to make reductions in emissions
cheaply and with minimal economic disruption. Whether or not this will
happen is impossible to know until we try.
521

14.4.3 What can you do?
If you have decided that climate change is something that we need to
address, you may be wondering what you as an individual can do. There
certainly are personal choices you can make that will reduce your share of
emissions of greenhouse gases. Some of these choices will not only reduce
emissions but also save you money and benefit you in other ways (e.g.,
take public transportation or walk instead of driving, eat less meat). Other
choices may require upfront costs but will pay for themselves over
subsequent years (e.g., add insulation to your attic or switch to LED
lighting). Some choices are difficult to justify on economic grounds alone
(e.g., install photovoltaic solar panels on your house3).
You should certainly undertake all of the voluntary individual actions
that you can. But these individual actions are not going to lead to the
emissions reductions necessary to stabilize the climate. Those will require
collective, coordinated action at both the national and international levels.
That is why the single most important thing you can do is become
politically active – write letters to your representatives, participate in
rallies, talk to your friends and neighbors, and vote for politicians who
support action on climate.
522

14.5 A few final thoughts
In this book, I have tried to give a comprehensive overview of the climate
change problem. Unlike what you might hear in the public debate, much of
the science of climate change is extremely solid. There is no question that,
when you add a greenhouse gas to the atmosphere, the planet will warm
(Chapters 4 and 6). There is no question that human activities are
increasing the amount of greenhouse gas in our atmosphere (Chapter 5).
There is no question that the Earth is currently warming (Chapter 2), and it
is warming about as much as you would expect from the addition of
greenhouse gases (Chapter 7). This science is not new – much of it is a
century or more old and has stood the test of time.
Other aspects of the problem are less certain. Quantitative projections
of future climate change at regional scales still contain significant
uncertainty; uncertainties also exist in the emissions scenario the world
will follow (Chapter 8). Moreover, this uncertainty is magnified by
uncertainty in how this warming will impact humans and those aspects of
the environment that we care about. However, one conclusion is clear: If
climate change falls toward the upper end of the predicted range, we will
truly be remaking the face of the planet, and the results may be dire,
perhaps even catastrophic (Chapter 9).
We know how to solve this problem (Chapter 11, 12, and 14), but we
do not know how hard and expensive it will be – and we will not know
until we try. Paralyzed by this uncertainty, the world has made little
progress in solving this problem, despite decades of warnings from
scientists (Chapter 13).
523

I do not know what the future holds. But I do know that, if we are
going to navigate the coupled problems of energy and climate, we are
going to need people like you to get involved in all parts of the problem:
the political, the economic, and the scientific. Given the enormous
creativity and inventiveness of humans, there is no question that we can
solve the problem. I encourage you to get involved to ensure that we do.
524

Additional reading
N. Stern, The Economics of Climate Change: The Stern Review
(Cambridge: Cambridge University Press), 2007; W. Nordhaus, The
Climate Casino: Risk, Uncertainty, and Economics for a Warming World
(New Haven, CT: Yale University Press), 2013. These two economic
analyses compare the costs and benefits of action on climate change. Both
conclude that action is required, but they differ on how much action,
mainly as a result of differences in the discount rate. Stern advocates
strong action immediately, whereas Nordhaus advocates a slower ramping
down on emissions.
P. Krugman, “Building a Green Economy,” New York Times Magazine,
April 7, 2010. This offers a clear and concise summary of the economics
of climate change policy and a critique of cost-benefit calculations on
climate change (download at
www.nytimes.com/2010/04/11/magazine/11Economy-t.html).
A. E. Dessler and E. A. Parson, The Science and Politics of Global
Climate Change: A Guide to the Debate, 2nd ed. (Cambridge: Cambridge
University Press, 2010). Chapter 5 of that book covers the uncertainty
question; it also outlines in detail the elements of an effective international
response to climate change as well as how we might get there.
S. M. Gardiner, A Perfect Moral Storm: The Ethical Tragedy of Climate
Change (Oxford: Oxford University Press, 2013). This is an accessible
primer describing the ethical issues of the climate change problem.
525

Problems
1. What is the single most important thing the world needs to do to
address climate change? Why?
2. A friend argues, “We must be certain climate change is a problem
before we take action.” Another friend argues, “We must take action
if the slightest chance exists that climate change could be
catastrophic.” How do you determine which one is right? Which one
(if either) do you judge to be correct?
3. How does irreversibility of the choices affect policy decisions?
4. Imagine that your hometown is always at risk of being destroyed
by some natural disaster (tornado, hurricane, earthquake). How much
would you pay each year to eliminate the chance that the disaster
would occur in that year? What are some ways you could determine
this value?
5.
a) Explain conceptually the role that discounting plays in
determining climate change policy.
b) If you change the discount rate from 0 percent to 4 percent, how
will this change your policy?
6. Juries in criminal trials are given a standard of evidence that must
be crossed in order to find a defendant guilty. What is it? What would
the standard be if our society decided that the worse error is to acquit
a guilty person?
526

7. For climate impacts happening in fifty years, how much effort
should we make to eliminate those? What about impacts happening in
500 years?
8. Economic analyses struggle to assign a monetary value to a small
chance of a truly terrible outcome. To see this, imagine that some
otherwise unavoidable activity carries with it a 0.1 percent chance of
killing you. How much money would you spend to avoid that
activity? What if the risk of death were 1 percent or 10 percent?
1 Quoted in R. Susskind, The One Percent Doctrine: Deep Inside
America’s Pursuit of Its Enemies since 9/11 (New York: Simon &
Schuster, 2006).
2 See Nordhaus (2008), p. 22.
3 This might make financial sense if you receive enough of a subsidy
from the government.
527

References
Balmaseda, M. A., K. E. Trenberth, and E. Kaellen, “Distinctive Climate
Signals in Reanalysis of Global Ocean Heat Content,” Geophysical
Research Letters, 40 (2013), 1754–1759, doi: 10.1002/grl.50382.
Ciais, P., C. Sabine, G. Bala, L. Bopp, V. Brovkin, J. Canadell, A.
Chhabra, R. DeFries, J. Galloway, M. Heimann, C. Jones, C. Le Quéré, R.
B. Myneni, S. Piao, and P. Thornton, “Carbon and Other Biogeochemical
Cycles,” in T. F. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S. K. Allen, J.
Boschung, A. Nauels, Y. Xia, V. Bex, and P. M. Midgley (eds.), Climate
Change 2013: The Physical Science Basis. Contribution of Working
Group I to the Fifth Assessment Report of the Intergovernmental Panel on
Climate Change (Cambridge and New York: Cambridge University Press,
2013), pp. 465–570, doi: 10.1017/CBO9781107415324.015.
Collins, M., R. Knutti, J. Arblaster, J.-L. Dufresne, T. Fichefet, P.
Friedlingstein, X. Gao, W. J. Gutowski, T. Johns, G. Krinner, M.
Shongwe, C. Tebaldi, A. J. Weaver, and M. Wehner, “Long-Term Climate
Change: Projections, Commitments and Irreversibility,” in T. F. Stocker,
D. Qin, G.-K. Plattner, M. Tignor, S. K. Allen, J. Boschung, A. Nauels, Y.
Xia, V. Bex, and P. M. Midgley (eds.), Climate Change 2013: The
Physical Science Basis. Contribution of Working Group I to the Fifth
Assessment Report of the Intergovernmental Panel on Climate Change
(Cambridge and New York: Cambridge University Press), pp. 1029–1136,
doi: 10.1017/CBO9781107415324.024.
528

Deffeyes, K. S., Beyond Oil: The View from Hubbert’s Peak (New York:
Hill and Wang, 2006).
Denman, K. L., G. Brasseur, A. Chidthaisong, P. Ciais, P. M. Cox, R. E.
Dickinson, D. Hauglustaine, C. Heinze, E. Holland, D. Jacob, U.
Lohmann, S. Ramachandran, P. L. da Silva Dias, S. C. Wofsy, and X.
Zhang, “Couplings between Changes in the Climate System and
Biogeochemistry,” in S. Solomon, D. Qin, M. Manning, Z. Chen, M.
Marquis, K. B. Averyt, M. Tignor, and H. L. Miller (eds.), Climate
Change 2007: The Physical Science Basis. Contribution of Working
Group I to the Fourth Assessment Report of the Intergovernmental Panel
on Climate Change (Cambridge: Cambridge University Press, 2007),
499–588.
Dessler, A. E., and E. A. Parson, The Science and Politics of Global
Climate Change: A Guide to the Debate, 2nd ed. (Cambridge: Cambridge
University Press, 2010).
Forster, P., V. Ramaswamy, P. Artaxo, T. Berntsen, R. Betts, D. W. Fahey,
J. Hay-wood, J. Lean, D. C. Lowe, G. Myhre, J. Nganga, R. Prinn, G.
Raga, M. Schulz, and R. Van Dorland, “Changes in Atmospheric
Constituents and in Radiative Forcing,” in S. Solomon, D. Qin, M.
Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor, and H. L. Miller
(eds.), Climate Change 2007: The Physical Science Basis. Contribution of
Working Group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change (Cambridge: Cambridge
University Press, 2007), 129–234.
Hansen, J., R. Ruedy, M. Sato, and K. Lo, “Global Surface Temperature
Change,” Reviews of Geophysics, 48 (2010) RG4004, doi:
10.1029/2010RG000345.
529

Houghton, J., “Madrid 1995: Diagnosing Climate Change,” Nature 455
(2008), 737–738 (accessible online at
http://www.nature.com/nature/journal/v455/n7214/full/455737a.html).
IPCC, “Summary for Policymakers,” in S. Solomon, D. Qin, M. Manning,
Z. Chen, M. Marquis, K. B. Averyt, M. Tignor, and H. L. Miller (eds.),
Climate Change 2007: The Physical Science Basis. Contribution of
Working Group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change (Cambridge: Cambridge
University Press, 2007a), 1–18.
IPCC, “Summary for Policymakers,” in B. Metz, O. R. Davidson, P. R.
Bosch, R. Dave, and L. A. Meyer (eds.), Climate Change 2007:
Mitigation. Contribution of Working Group III to the Fourth Assessment
Report of the Intergovernmental Panel on Climate Change (Cambridge:
Cambridge University Press, 2007b), 1–24.
IPCC, “Summary for Policymakers,” in T. F. Stocker, D. Qin, G.-K.
Plattner, M. Tignor, S. K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex,
and P. M. Midgley (eds.), Climate Change 2013: The Physical Science
Basis. Contribution of Working Group I to the Fifth Assessment Report of
the Intergovernmental Panel on Climate Change (Cambridge and New
York: Cambridge University Press, 2013), 3–32.
Jansen, E., J. Overpeck, K. R. Briffa, J.-C. Duplessy, F. Joos, V. Masson-
Delmotte, D. Olago, B. Otto-Bliesner, W. R. Peltier, S. Rahmstorf, R.
Ramesh, D. Raynaud, D. Rind, O. Solomina, R. Villalba, and D. Zhang,
“Palaeoclimate,” in S. Solomon, D. Qin, M. Manning, Z. Chen, M.
Marquis, K. B. Averyt, M. Tignor, and H. L. Miller (eds.), Climate
Change 2007: The Physical Science Basis. Contribution of Working
Group I to the Fourth Assessment Report of the Intergovernmental Panel
530

http://www.nature.com/nature/journal/v455/n7214/full/455737a.html

on Climate Change (Cambridge: Cambridge University Press, 2007),
433–498.
Kirtman, B., S. B. Power, J. A. Adedoyin, G. J. Boer, R. Bojariu, I.
Camilloni, F. J. Doblas-Reyes, A. M. Fiore, M. Kimoto, G. A. Meehl, M.
Prather, A. Sarr, C. Schär, R. Sutton, G. J. van Oldenborgh, G. Vecchi, and
H. J. Wang, “Near-term Climate Change: Projections and Predictability,”
in T. F. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S. K. Allen, J.
Boschung, A. Nauels, Y. Xia, V. Bex, and P. M. Midgley (eds.), Climate
Change 2013: The Physical Science Basis. Contribution of Working
Group I to the Fifth Assessment Report of the Intergovernmental Panel on
Climate Change (Cambridge and New York: Cambridge University Press,
2013), 953–1028.
Lemke, P., J. Ren, R. B. Alley, I. Allison, J. Carrasco, G. Flato, Y. Fujii,
G. Kaser, P. Mote, R. H. Thomas, and T. Zhang, “Observations: Changes
in Snow, Ice and Frozen Ground,” in S. Solomon, D. Qin, M. Manning, Z.
Chen, M. Marquis, K. B. Averyt, M. Tignor, and H. L. Miller (eds.),
Climate Change 2007: The Physical Science Basis. Contribution of
Working Group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change (Cambridge: Cambridge
University Press, 2007), 377–384.
Lisiecki, L. E., and M. E. Raymo, “A Pliocene-Pleistocene Stack of 57
Globally Distributed benthic delta O-18 Records,” Paleoceanography 20
(2005), PA1003, doi: 10.1029/2004PA001071.
Marcott, S. A., J. D. Shakun, P. U. Clark, and A. C. Mix, “A
Reconstruction of Regional and Global Temperature for the Past 11,300
Years,” Science 339 (2013), 1198–1201, doi: 10.1126/science.1228026.
Meehl, G. A., C. Tebaldi, G. Walton, D. Easterling, and L. McDaniel,
531

“Relative Increase of Record High Maximum Temperatures Compared to
Record Low Minimum Temperatures in the U.S.,” Geophysical Research
Letters 36 (2009) L23701, doi: 10.1029/2009GL040736.
Myhre, G., D. Shindell, F.-M. Bréon, W. Collins, J. Fuglestvedt, J. Huang,
D. Koch, J.-F. Lamarque, D. Lee, B. Mendoza, T. Nakajima, A. Robock,
G. Stephens, T. Takemura, and H. Zhang, “Anthropogenic and Natural
Radiative Forcing,” in T. F. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S.
K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex, and P. M. Midgley
(eds.), Climate Change 2013: The Physical Science Basis. Contribution of
Working Group I to the Fifth Assessment Report of the Intergovernmental
Panel on Climate Change (Cambridge and New York: Cambridge
University Press, 2013), 659–740.
National Research Council, Surface Temperature Reconstructions for the
Last 2,000 Years. Board on Atmospheric Sciences and Climate
(Washington, DC: National Academies Press, 2006).
Nerem, R. S., D. P. Chambers, C. Choe, and G. T. Mitchum, “Estimating
Mean Sea Level Change from the TOPEX and Jason Altimeter Missions,”
Marine Geodesy, 33 (sup. 1) (2010), 435–446.
Nordhaus, W., A Question of Balance: Weighing the Options on Global
Warming Policies (New Haven, CT: Yale University Press, 2008).
Petit, J. R., J. Jouzel, D. Raynaud, N. I. Barkov, J. M. Barnola, I. Basile,
M. Bender, J. Chappellaz, M. Davis, G. Delaygue, M. Delmotte, V. M.
Kotlyakov, M. Legrand, V. Y. Lipenkov, C. Lorius, L. Pepin, C. Ritz, E.
Saltzman, and M. Stievenard, “Climate and Atmospheric History of the
Past 420,000 Years from the Vostok Ice Core, Antarctica,” Nature 399
(1999), 429–436.
532

Royer, D. L., “CO2-Forced Climate Thresholds during the Phanerozoic,”
Geochimica et Cosmochimica Acta 70 (2006), 5665–5675.
Solomon, S., D. Qin, M. Manning, R. B. Alley, T. Berntsen, N. L. Bindoff,
Z. Chen, A. Chidthaisong, J. M. Gregory, G. C. Hegerl, M. Heimann, B.
Hewitson, B. J. Hoskins, F. Joos, J. Jouzel, V. Kattsov, U. Lohmann, T.
Matsuno, M. Molina, N. Nicholls, J. Overpeck, G. Raga, V. Ramaswamy,
J. Ren, M. Rusticucci, R. Somerville, T. F. Stocker, P. Whetton, R. A.
Wood, and D. Wratt, “Technical Summary,” in S. Solomon, D. Qin, M.
Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor, and H. L. Miller
(eds.), Climate Change 2007: The Physical Science Basis. Contribution of
Working Group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change (Cambridge: Cambridge
University Press, 2007), 19–92.
Solomon, S., G.-K. Plattner, R. Knutti, and P. Friedlingstein, “Irreversible
Climate Change due to Carbon Dioxide Emissions,” Proceedings of the
National Academy of Sciences 106 (2009), 1704–1709.
Stanton, E. A., and F. Ackerman, Florida and Climate Change: The Costs
of Inaction (Boston: Tufts University, 2007; available online at
http://ase.tufts.edu/gdae/Pubs/rp/FloridaClimate.html).
van Vuuren, D. P., et al., “The Representative Concentration Pathways: An
Overview,” Climatic Change 109 (2011), 5–31, doi: 10.1007/s10584-011-
0148-z.
Vaughan, D. G., J. C. Comiso, I. Allison, J. Carrasco, G. Kaser, R. Kwok,
P. Mote, T. Murray, F. Paul, J. Ren, E. Rignot, O. Solomina, K. Steffen,
and T. Zhang, “Observations: Cryosphere,” in T. F. Stocker, D. Qin, G.-K.
Plattner, M. Tignor, S. K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex,
and P. M. Midgley (eds.) Climate Change 2013: The Physical Science
533

http://ase.tufts.edu/gdae/Pubs/rp/FloridaClimate.html

Basis. Contribution of Working Group I to the Fifth Assessment Report of
the Intergovernmental Panel on Climate Change (Cambridge and New
York: Cambridge University Press, 2013).
Weart, S. R. The Discovery of Global Warming, 2nd ed. (Cambridge, MA:
Harvard University Press, 2008). Accessible online at
www.aip.org/history/climate/index.htm
Zachos, J., M. Pagani, L. Sloan, E. Thomas, and K. Billups, “Trends,
Rhythms, and Aberrations in Global Climate 65 Ma to Present,” Science
292 (2001), 686–693.
534

http://www.aip.org/history/climate/index.htm

Index
2°C target, 238
abrupt climate change, 158
acid rain, 77, 216
adaptation, 178
definition of successful, 181
role of government, 179
role of wealth, 180
additionality, 205
aerosols, 95, 220
black carbon, 96, 100
direct radiative effect, 97
indirect effect, 97, 99
mineral dust, 96
residence time, 95, 99
sulfate, 95
volcanoes, 95
affluence, 127, 131
Agassiz, Louis, 211
agriculture, 153
albedo, 55
annual cycle, carbon dioxide, 69
anomalies, 17
Arrhenius, Svante, 211
535

atmosphere, properties of, 57
atmospheric composition, 67
argon, 67
greenhouse gases. See greenhouse gases
nitrogen, 67
oxygen, 67, 71, 79
water vapor, 68
Australia, climate policy, 227
biomass energy, 185
density, 185
blackbody radiation, 41
boundary value problems, 140
Bush, George H.W., 217, 224
Bush, George W., 224
emissions target, 225
Callendar, Guy Stewart, 212
calorie, 38
Canada, climate policy, 227
cap and trade, 201, 217
additionality, 205
comparison to carbon tax, 203
escape valve, 204
offsets, 205
permit issuance, 202
carbohydrate, 70
carbon capture and storage, 187, 192
carbon cycle, 67
carbon dioxide, 68, 69
annual cycle, 69
536

association with temperature, 118, 119
carbon-14 half-life, 82
isotopes, 82
lags temperature, 121
missing carbon, 79
radiative forcing, 94
radiocarbon dating, 83
radiocarbon dead, 83
year-to-year variability, 79
carbon dioxide, atmospheric increase, 77
carbon intensity, 129, 132
required reduction over 21st century, 240
carbon sequestration, 187
carbon tax, 198
additionality, 205
comparison to cap and trade, 203
offsets, 205
tax rate increase, 201
carbon-cycle engineering, 192
carbonic acid, 72
CCN. See cloud condensation nuclei
CCS. See carbon capture and storage
Celsius scale, 1
CFCs. See greenhouse gases, halocarbons
chasing water, 146
chemical weathering, 74
cherry picking, 23
China-U.S. emissions agreement, 228, 243
chlorofluorocarbons. See greenhouse gases, halocarbons
climate
definition, 1
537

precipitation, 3
temperature, 3
impacts. See impacts
climate change
definition, 4
climate sensitivity, 106
climate skeptics, 221
cloud condensation nuclei, 97
cloud feedback, 103
cloud particle size, 97
committed warming, 92
common but differentiated responsibilities, 222, 224
Concorde, 216
consensus, 9
conservation of energy, 46
continental drift. See plate tectonics
conventional regulations, 197, 200
Copenhagen Accord, 226
dangerous climate change, 222, 236
D-Day, 2
decisions under uncertainty
criminal trial, 234
irreversibility, 235
national defense, 234
deep ocean, 72
deforestation, 77
Dessler, Kasper, 45
discount rate, 169
growth discounting, 171
social cost of carbon, 172
538

time discounting, 170
discounting, 169, 237
doubling time, 163
drought, 152
Earth’s orbit, 114
eccentricity, 114
ecosystem services, 156
El Nino, 22, 117
electromagnetic radiation, 39
emissions scenarios, 133
emissions spectra, 41
incandescent light bulb, 43
Sun, 43
energy, 38
energy balance, 46
energy intensity, 128, 132
ENSO. See El Nino
Eocene Climatic Optimum, 31
equator, 5
error bars, 20
escalator plot, 23
escape value, 204
ETS. See EU Emissions Trading System
EU Emissions Trading System, 206
expert witness, 8
expertice, 8
consensus, 9, 11
trustworthiness, 11
exponential growth, 162
externality, 189
539

Fahrenheit scale, 1
FCCC. See Framework Convention on Climate Change
feedbacks, 102
carbon cycle, 105, 121
clouds, 103
fast, 104
ice-albedo, 103
lapse rate, 103
negative, 103
numeric values, 106
positive, 103
slow, 104, 108
vegetation, 105
water vapor, 103
weathering thermostat, 105
flexibility, 201
floods, 152
fossil fuels, 75, 76
limits, 138
Fourier, Joseph, 56, 211
Framework Convention on Climate Change, 222
common but differentiated responsibilities, 222, 224
dangerous climate change, 222, 236
emissions target, 223
goal, 222
historical responsibility, 223
precautionary principle, 223
free market and climate, 188
GDP. See gross domestic product
geoengineering
540

carbon-cycle engineering, 192
political issues, 191
solar radiation management, 190
geongineering, 190
gigatonne of carbon, 71
glaciations, 31
glaciers, loss of, 24
global cooling, predictions of, 220
greenhouse effect, 57
n-layer model, 59
one-layer model, 57
two-layer model, 58
greenhouse gas intensity, 128
greenhouse gases, 67
carbon dioxide, 68, 69
cause of present warming, 117
halocarbons, 69, 94, 216, 218
methane, 68, 69, 85, 94
nitrous oxide, 68, 94
ozone, 68, 94
stratospheric water vapor, 95
water vapor, 68
gross domestic product, 127
growth discounting, 171
GtC. See gigatonne of carbon
halocarbons, 69, 94
Hansen, James, 221
heat waves, 151
hiatus in warming, 22
Holocene, climate of, 33
541

hurricanes, 152
hydrochlorofluorocarbons. See greenhouse gases, halocarbons
hydroelectric energy, 186
ice ages, 32, 115, 145
ice sheets, loss of, 25
ice-albedo feedback, 103
impacts
abrupt changes, 158
agriculture, 153
drought, 152
extinction of species, 155
extreme weather, 151
floods, 152
health, 157
heat, 3
heat waves, 151
hurricanes, 152
national security, 157
ocean acidification, 150, 155
precipitation, 149, 152
rate of warming, 156
sea level rise, 150, 154
temperature, 147
water availability, 153
incandescent light bulb, 43
industrial revolution, 77
information and voluntary methods, 207
infrared, 40
initial value problems, 140
interest rate, 162
542

interglacials, 32
Intergovernmental Panel on Climate Change. See IPCC
internal energy, 39
internal variability, 116
International Geophysical Year, 214
international negotiations, 241
IPAT relation, 128
IPCC
alternative sources of information, 12
formation, 221
process, 9
Summary for Policymakers, 9
isotopes, 82
Joule, 38
Kaya Identity. See IPAT relation
Keeling, Charles D., 79
Keeling, Ralph, 79
Kelvin scale, 39
killer smog of London, 213
Kuhn, Thomas, 119
Kyoto Protocol, 223
emissions target, 224
entry into force, 225
final results, 228
flexibility, 224
La Nina, 117
land biosphere, 71
land-use changes
carbon emissions, 77
543

radiative forcing, 100
lapse-rate feedback, 103
latitude, 5
solar energy distribution, 55
lifetime. See turnover time
Little Ice Age, 33
Little Red Riding Hood, 212
longitude, 5
Malthus, Thomas, 168
Malthusian catastrophe, 168
marginal cost, 198
market failure, 189
market-based regulations, 198
carbon tax, 198
Mars, 62
McCain, John, 226
Medieval Warm Period, 34
Mercury, 62
methane, 68, 69, 85, 94
micron, 40
microwave oven, 48
microwaves, 40
mid-latitudes, 13
Milankovitch cycles, 115
mitigation, 182
required cuts in emissions, 182
role of technology, 183
mixed layer, 72
Montreal Protocol, 218, 222
544

national emissions targets, 241
negative feedback, 103
Nielsen-Gammon, John, 83
nitrous oxide, 68, 94
Nordhaus, William, 243
nuclear energy, 186
nuclear weapons, 213
Obama, Barack, 226
obliquity, 115
ocean acidification, 72, 120, 150, 155
ocean carbon reservoir, 71
ocean heat content, 26
offsets, 205
opinion leader, 8
orbit of the Earth, 114
outlier, 118
oven, 47
cooking a turkey, 48
oven, microwave, 48
oxygen, 67, 71, 79
ozone, 68, 94
ozone depletion, 218
ozone hole, 218
Paleocene-Eocene Thermal Maximum, 31, 120, 158
paleoproxies, 29
glacier, 29
ice cores, 30
ocean sediments, 30
tree rings, 30
545

parts per million, 68
comparisons to percent, 68
percent, 68
perihelion, 115
permafrost, 71, 104
PETM. See Paleocene-Eocene Thermal Maximum
photons, 40
photosynthesis, 70, 71
Planets, climate of, 61
plate tectonics, 75, 112
Indian subcontinent collision, 76, 113
polar regions, 13
population, 127, 130
positive feedback, 103
power, 38
precautionary principle, 223
predictability of the climate, 139
present value, 169
prime meridion, 5
radiative forcing, 92
radiocarbon dating, 83
radiocarbon dead, 83
RCPs. See Representative Concentration Pathways
regional emissions reduction systems, 225
Regional Greenhouse Gas Initiative, 225
renewable energy, 184
Representative Concentration Pathways, 133
radiative forcing, 135
temperature projections, 136
residence time. See turnover time
546

respiration, 70, 71
RF. See radiative forcing
RGGI, 225
rock carbon reservoir, 74
rule of 72, 163
satellite thermometer record, 21
adjustments and issues, 21
comparison to surface thermometers, 22
sea ice, loss of, 25
sea level rise, 27, 145, 150, 154
consistency with other measurements, 27
shadow area, 53
ship tracks, 98
skeptics, 221
snowball Earth, 31
social cost of carbon, 172
solar constant, 52
variability, 100, 113
solar energy, 184
density, 184
falling on Earth, 53
solar photovoltaic, 184
solar radiation management, 190
solar thermal, 184
Stefan-Boltzmann
constant, 45
equation, 44
stratospheric water vapor, 95
Summary for Policymakers, 9
Sun
547

emissions spectra, 43
supersonic airliner, 215
surface thermometer record, 19
adjustments, 20
comparison to satellite record, 22
warming has stopped, 22
Tea Party, 227
temperature, 39
Celsius, 1
Fahrenheit, 1
Kelvin, 39
temperature anomalies, 17
The Structure of Scientific Revolution, 119
thermometers. See surface thermometer record
time discounting, 170
time lags in climate system, 92
time scale, 74
tobacco debate, 10
tobacco strategy, 214
tragedy of the commons, 189
tropics, 13
turnover time, 73
Tyndall, John, 211
U.S.-China emissions agreement, 228, 243
ultraviolet, 40
Venus, 62
solar constant, 52
visible photons, 40
volcanoes, 74, 95
548

Volcanoes, 22, 30, 75
warming
observed, 19
predicted, 136
predicted distribution, 147
water availability, 153
water vapor feedback, 103
watts, 38
wavelength, 40
infrared, 40
ultraviolet, 40
visible, 40
weather
definition, 1
Wien’s displacement law, 41
wind energy, 184
density, 185
Year without a summer, 96
Younger Dryas, 158
549

Half title
Title page
Imprints page
Dedication
Contents
Preface
Acknowledgments
1 An introduction to the climate problem
1.1 What is climate?
1.2 What is climate change?
1.3 A coordinate system for the Earth
1.4 Why you should believe this textbook
1.5 Chapter summary
Additional reading
Terms
Problems
2 Is the climate changing?
2.1 Recent climate change
2.1.1 Surface thermometer record
2.1.2 Satellite measurements of temperature
2.1.3 Ice
2.1.3.1 Glaciers
2.1.3.2 Sea ice
2.1.3.3 Ice sheets
2.1.4 Ocean temperatures
2.1.5 Sea level
2.1.6 Putting it all together: Is today’s climate changing?
2.1.7 What is not evidence of climate change
2.2 Climate over the Earth’s history
2.2.1 Paleoproxies
2.2.2 The Earth’s long-term climate record
2.3 Chapter summary
Additional reading
Terms
Problems
3 Radiation and energy balance
3.1 Temperature and energy
3.2 Electromagnetic radiation
3.3 Blackbody radiation
3.4 Energy balance
3.5 Chapter summary
Additional reading
Terms
Problems
4 A simple climate model
4.1 The source of energy for our climate system
4.2 Energy loss to space
4.3 The greenhouse effect
4.3.1 One-layer model
4.3.2 Two-layer model
4.3.3 n-layer model
4.4 Testing our theory with other planets
4.5 Chapter summary
Additional reading
Terms
Problems
5 The carbon cycle
5.1 Greenhouse gases and our atmosphere’s composition
5.2 Atmosphere-land biosphere-ocean carbon exchange
5.2.1 Atmosphere-land biosphere exchange
5.2.2 Atmosphere-ocean carbon exchange
5.2.3 The combined atmosphere-land biosphere-ocean system
5.3 Atmosphere-rock exchange
5.4 How are humans perturbing the carbon cycle?
5.5 Some commonly asked questions about the carbon cycle
5.6 The long-term fate of carbon dioxide
5.7 Methane
5.8 Chapter summary
Additional reading
Terms
Problems
6 Forcing, feedbacks, and climate sensitivity
6.1 Time lags in the climate system
6.2 Radiative forcing
6.2.1 Greenhouse gases
6.2.2 Aerosols
6.2.3 Land-use changes
6.2.4 Changes in the Sun
6.2.5 Total net forcing
6.3 Feedbacks
6.3.1 Fast feedbacks
6.3.2 Slow feedbacks
6.4 Climate sensitivity
6.4.1 Feedback math
6.4.2 Sensitivity
6.5 Chapter summary
Additional reading
Terms
Problems
7 Why is the climate changing?
7.1 The first suspect: Continental drift
7.2 The Sun
7.3 The Earth’s orbit
7.4 Internal variability
7.5 Greenhouse gases
7.6 Putting it all together
7.7 Chapter summary
Terms
Additional reading
Problems
8 Predictions of future climate change
8.1 The factors that control emissions
8.2 How these factors have changed in the recent past and how will they change in the future
8.2.1 Population
8.2.2 Affluence
8.2.3 Technology
8.3 Emissions scenarios
8.4 Predictions of future radiative forcing
8.5 Predictions of future climate
8.5.1 Over the next century
8.5.2 Climate change beyond 2100
8.6 Is the climate predictable?
8.7 Chapter summary
Additional reading
Terms
Problems
9 Impacts of climate change
9.1 Why should you care about climate change?
9.2 Physical impacts
9.2.1 Temperature
9.2.2 Precipitation
9.2.3 Sea-level rise and ocean acidification
9.2.4 Loss of ice
9.2.5 Extreme events
9.3 Impacts of these changes
9.4 Abrupt climate changes
9.5 Chapter summary
Additional reading
Terms
Problems
10 Exponential growth
10.1 What is exponential growth?
10.2 The rule of 72
10.3 Limits to exponential growth
10.4 Discounting
10.4.1 The time value of money
10.4.2 The discount rate
10.5 Putting it together: The social cost of carbon
10.6 Chapter summary
Additional reading
Terms
Problems
11 Fundamentals of climate change policy
11.1 Adaptation
11.2 Mitigation
11.2.1 Technologies to reduce carbon intensity
11.2.2 Policies to reduce carbon emissions
11.3 Geoengineering
11.4 Chapter summary
Additional reading
Terms
Problems
12 Mitigation policies
12.1 Conventional regulations
12.2 Market-based regulations
12.2.1 Carbon tax
12.2.2 Cap and trade
12.2.3 Carbon tax versus cap and trade
12.2.4 Offsets
12.3 Information and voluntary methods
12.4 Chapter summary
Terms
Additional reading
Problems
13 A brief history of climate science and politics
13.1 The beginning of climate science
13.2 The emergence of environmentalism
13.2.1 The tobacco strategy
13.3 The 1970s and 1980s: Supersonic airliners, acid rain, and ozone depletion
13.4 The year everything changed: 1988
13.5 The Framework Convention on Climate Change: The First Climate Treaty
13.6 The Kyoto Protocol
13.7 The George W. Bush years: 2001–2008
13.8 The Obama years: 2009–today
13.9 The breakthrough: U.S.-China bilateral agreement
13.10 Chapter summary
Terms
Additional reading
Problems
14 Putting it together: A long-term policy to address climate change
14.1 What makes climate change such a difficult problem?
14.2 Decisions under uncertainty: Should we reduce emissions?
14.3 Picking a long-term goal
14.3.1 Cost versus benefits
14.3.2 Target: 2°C
14.4 How do we get there?
14.4.1 The physics of a 2°C limit
14.4.2 How to get there from here
14.4.3 What policies should look like
14.4.3 What can you do?
14.5 A few final thoughts
Additional reading
Problems
References
Index