Earth’s Climate System

This chapter provides an overview of the factors that influence Earth’s climate. The relation between reconstructions of global mean surface temperature and estimates of atmospheric carbon dioxide (CO2) over the past 500 million years is first described. Vast variations in climate on geologic time scales, driven by natural fluctuations of CO2, are readily apparent. We then shift attention to the time period 1765 to present, known as the Anthropocene, during which human activity has strongly influenced atmospheric CO2, other greenhouse gases (GHGs), and Earth’s climate. Two mathematical concepts essential for quantitative understanding of climate change, radiative forcing and global warming potential, are described. Next, fingerprints of the impact of human activity on rising temperature and the abundance of various GHGs over the course of the Anthropocene are presented. We conclude by showing Earth is in the midst of a remarkable transformation. In the past, radiative forcing of climate represented a balance between warming due to rising GHGs and cooling due to the presence of suspended particles (aerosols) in the troposphere. There presently exists considerable uncertainty in the actual magnitude of radiative forcing of climate due to tropospheric aerosols, which has important consequences for our understanding of the climate system. In the future, climate will be driven mainly by GHG warming because aerosol precursors are being effectively removed from pollution sources, due to air quality legislation enacted in response to public health concerns.

thrived in an oxygen (O 2 ) free environment. These organisms had no nucleus and reproduced by cell division. The first prokaryotes likely made organic matter by combining carbon dioxide (CO 2 ) with molecules such as hydrogen sulfide (H 2 S), releasing water (H 2 O) and elemental sulfur to the environment (Canfield and Raiswell 1999).
Early in Earth's history the favored atmospheric fate for carbon-bearing compounds was methane (CH 4 ), because the atmosphere was in a state chemists call "reducing". Stellar astronomy indicates that at the time early life formed, the luminosity of our Sun was about 30 % less than today, which should have caused ancient oceans to freeze. As explored in detail throughout this book, CH 4 is a more potent greenhouse gas (GHG) than CO 2 . Extremely high levels of atmospheric CH 4 and ammonia (NH 3 ), another reduced compound that is also a strong GHG, were likely responsible for preventing Earth's ancient oceans from freezing (Sagan and Mullen 1972).
Prokaryotes were the first to develop photosynthesis, the ability to convert sunlight, carbon dioxide (CO 2 ), and water (H 2 O) into glucose C 6 H 12 O 6. Eventually, prokaryotic photosynthesis caused atmospheric O 2 to rise from about one part per million of all air molecules to 21 %. Margulis and Sagan (1986) call the initial buildup of atmospheric O 2 the greatest environmental crisis Earth has ever endured. At the time, O 2 was toxic to most life on Earth. As a result, a mass extinction called the Great Oxygenation Event occurred about 2.5 Bybp. One can only imagine the emergency meetings of bacterial communities, seeking to ban their photosynthetic cousins in an effort to halt the build-up of atmospheric O 2 .
The rise of atmospheric O 2 had enormous consequences. For the first time in Earth's history, CO 2 was the favored state for atmospheric carbon gases. Conversion of atmospheric CH 4 to CO 2 likely led to Earth's first glaciation event about 2.4 Bybp (Frei et al. 2009). The build-up of O 2 also led to formation of Earth's protective ozone (O 3 ) layer, which was necessary for life to emerge from sea to land. Finally, the global, atmospheric chemical shock induced by the Great Oxygenation Event facilitated the evolution of eukaryotes: nucleated cells that metabolize O 2 . You are made of eukaryotic cells! Plant life first appeared on land about 500 million years before present (Mybp) (Kenrick and Crane 1997). Even though, as alluded to above, much is known about climate and the state of Earth's atmosphere prior to this time, reconstructions of global variations in Earth's climate and atmospheric CO 2 are only available for the most recent 500 million years. Figure 1.1 shows the variations in the global mean surface temperature and the abundance of carbon dioxide (CO 2 ) over the past 500 million years. The temperature estimates are anomalies (ΔT) with respect to the mean state of Earth's climate that existed during recent pre-industrial time (i.e., years 1850-1900). Notable events regarding the evolution of life (Dinosaurs, Rise of Mammals, etc.) are indicated to the left and phenomenon regarding the global carbon cycle and climate (Rise of Forests, Greenland Glaciation, etc.) are marked to the right. This figure is our composite of a considerable number of paleoclimatic studies, as described in Methods. To facilitate discussion of this figure, six Eras are denoted. These should not be confused with the formal use of the word Era by Geologists. Each era in Fig. 1.1   .1 shows that Earth has undergone vast changes in climate as well as the abundance of atmospheric CO 2 . A logarithmic scale is used to represent the mixing ratio of atmospheric CO 2 , because the radiative forcing of climate (see Sect. 1.2.1) is proportional to the logarithm of CO 2 . Figure 1.1 shows, clearly and beyond debate, the strong association of Earth's climate and atmospheric CO 2 . Looking backwards in time, as CO 2 rises, Earth warms. On geological time scales, atmospheric CO 2 is controlled by the carbonate-silicate cycle (Berner et al. 1983). Atmospheric supply of CO 2 occurs during volcanic eruptions and hydrothermal venting. Atmospheric removal of CO 2 is more complicated. The weathering of minerals converts atmospheric CO 2 into a water soluble form of carbon; ocean organisms incorporate soluble carbon into their shells and, when these animals perish, their shells sink to the ocean floor. Plate tectonics buries the sinking sediment, after which the carbon either remains in Earth's mantle or, on occasion, is spewed back to the atmosphere-ocean system, via either volcanoes or deep sea vents.
The first dramatic perturbation to the carbonate-silicate cycle was induced by the rise of forests. About 500 Mybp, atmospheric CO 2 may have been as high as 5000 parts per million by volume (ppm), more than a factor of 10 larger than today ( Fig. 1.1) (Berner 1997). Of course, Earth was also exceedingly warm compared to today. The first plants to evolve, bryophytes, were algae-like organisms that probably eased the transition from sea to land by finding homes on moist rocky surfaces. Bryophytes are known as non-vascular plants; they lack roots to transport moisture. Moss is a modernday bryophyte. Vascular pteridophytes (fern-like organisms) evolved about 500 Mybp, soon leading to the rise of forests. This resulted in a steady, dramatic decline in atmospheric CO 2 because early forests lacked the abundant bacteria, fungi, and small soil animals that recycle plant matter in contemporary forests. Carbon in the forests that prevailed during the time depicted as Era 6 of Fig. 1.1 was buried and converted to modern day coal and natural gas deposits, due to the intense heat and pressure within Earth's mantle. This carbon is now being released back to the atmosphere-ocean system, perhaps to generate the electricity used to help you read this book.
The next event that transformed the global carbon cycle was the rise of the Himalayas (Raymo and Ruddiman 1992). During the period of time depicted in Era 5, plate tectonics resulted in the formation of the modern-day continents. The exposure of fresh minerals due to the vast tectonic activity associated with formation of the Himalayan mountain range, the largest in the world, led to the steady draw down of CO 2 and associated cooling depicted in Era 5 of Fig. 1.1.
About 3 Mybp, two remarkable events occurred. Our predecessor Lucy (Aus tralopithecus afarensis) roamed modern day Ethiopia (Johanson and White 1979). At about the same time, Greenland first became glaciated (Lunt et al. 2008). While the emergence of an early human ancestor who walked in an upright manner is in no way related to the glaciation of Greenland, it is worth noting that global mean surface temperature and atmospheric CO 2 at the time of Lucy were both estimated to be at modern, pre-industrial levels (Era 4, Fig. 1.1).
The most compelling association of CO 2 and climate is provided by the covariance of these quantities during the past 800,000 years (Era 3, Fig. 1.1) (Imbrie and Imbrie 1979). Earth's climate oscillated between glaciated and inter-glacial states, with atmospheric CO 2 levels of about 200 ppm and 280 ppm, respectively, characterizing each state (Barnola et al. 1987). Lower levels of atmospheric CO 2 prevailed during glacial times due to more productive ocean biogeochemical uptake (Marino et al. 1992), perhaps facilitated by the oceanic supply of iron resulting from the grinding of glaciers on rock (Martin 1990). The ultimate pace-maker of these cycles is orbital variations of Earth about the Sun, known as Milankovitch cycles (Imbrie and Imbrie 1979).
The precise timing of the rise and fall of temperature and CO 2 during Era 3 of Fig. 1.1 is a source of considerable dispute between the climate "believers" and "deniers". The initial, literal interpretation of the ice core record suggested that changes in temperature proceeded variations in CO 2 by about 800 years (Caillon et al. 2003). If so, the deniers argue, then CO 2 is responding to, rather than driving, global climate change. It is essential to appreciate that: the ice core record of CO 2 is discerned by measuring the composition of bubbles trapped in ice; historic temperature is quantified by measuring isotopic composition of the hydrogen and/or oxygen elements within the ice; and bubbles within the sampled ice cores move with respect to the surrounding ice over geologic time. A recent re-analysis of the timing of variations in temperature and CO 2 of an Antarctic ice core, which considers movement of bubbles with respect to the surrounding ice, reveals synchronous variation within the uncertainty of measurement (Parrenin et al. 2013). This interpretation supports the view that changes in atmospheric CO 2 did indeed drive glacial/interglacial transitions.
Ancient air preserved in ice cores reveals that when Earth underwent glacial conditions during Era 3 of Fig. 1.1, atmospheric CH 4 fell to about 0.4 ppm. During interglacial periods, atmospheric CH 4 reached a value of 0.7 ppm (Petit et al. 1999;Loulergue et al. 2008). Natural sources of CH 4 vary by an amount large enough to induce considerable variations in atmospheric abundance, with some consequence for the radiative forcing of climate. Methane is released to the atmosphere when frequently flooded regions (wetlands) experience low oxygen (anaerobic) conditions. It is likely that the higher levels of CH 4 during warm epochs was due to a larger preponderance of wetlands, particularly in the northern hemisphere (NH), as these regions went from ice-covered to ice-free as Earth transitioned from glacial to interglacial conditions (Brook et al. 2000;Sowers 2006). Variations in atmospheric CH 4 also played a role in driving glacial/interglacial climate cycles.
The correlation of temperature and atmospheric CO 2 over vast periods of Earth's history is firmly established by the hundreds of studies that have led to our composite Fig. 1.1. Of course, correlation does not necessarily imply causation. The radiative forcing of climate due to CO 2 and other GHGs is explored in great detail later in this chapter.
Modern Homo sapiens evolved about 200,000 years ago and left Africa about 100,000 years ago (Carto et al. 2009). The paths of early humans were influenced by various rapid climate change events that took place at the end of Era 3 and the start of Era 2 of Fig. 1.1. During the height of most recent glaciation, about 20,000 years ago, modern day Manhattan was under a sheet of ice nearly half a mile thick and global sea level was 120-130 m (about 400 ft!) lower than today. Human settlement of North America hugged the coastline, as the interior was inhospitable if not impassable. Scandinavia, England, Wales, Scotland, and Ireland were similarly buried under year-long ice.
Modern agriculture was invented during the Neolithic Revolution, about 12,500 years ago. The Earth was in the midst of climatic warming following the end of the last ice age, which has been implicated as a causal factor because the domestication of plants and animals occurred nearly simultaneously at places separated by great distance (Gepts and Papa 2001). Agriculture flourished, population grew, and humans colonized all parts of the Earth (except Antarctica) during the climatically quiescent times depicted at the end of Era 2 and start of Era 1 in Fig. 1.1.
Eventually, population rose to an extent that humans began to exert a measurable effect on the GHG levels of the atmosphere, a time period now known as the Anthropocene (Crutzen and Stoermer 2000). Figure 1.2 illustrates the time evolution of ΔT, CO 2 , CH 4 , nitrous oxide (N 2 O), and population during the past 2000 years, the so-called Common Era. The association of the rise in population and increased atmospheric levels of CO 2 , CH 4 , and N 2 O is again irrefutable. Before delving into the Anthropocene, we shall comment on one more controversy.
Europe experienced unusually warm temperatures from about 950-1250 AD, a time known as the Medieval Warm Period (MWP) (Moberg et al. 2005). Our Fig. 1.2, which relies on the global temperature reconstruction of Jones and Mann (2004), does not depict the MWP. The study of Jones and Mann (2004) suggests the MWP was regional in nature, with little to no global expression. The temperature record in Fig. 1.2, which has become known as "The Hockey Stick", has led to considerable controversy. One account is described in the book The Hockey Stick and the Climate Wars: Dispatches from the Front Lines (Mann 2012 Based on the analyses presented in the original papers by Mann et al. and this newer supporting evidence, the committee finds it plausible that the Northern Hemisphere was warmer during the last few decades of the 20th century than during any comparable period over the preceding millennium. The substantial uncertainties currently present in the quantitative assessment of large-scale surface temperature changes prior to about A.D. 1600 lower our confidence in this conclusion compared to the high level of confidence we place in the Little Ice Age cooling and 20th century warming. Even less confidence can be placed in the original conclusions by Mann et al. (1999) that 'the 1990s are likely the warmest decade, and 1998 the warmest year, in at least a millennium' because the uncertainties inherent in temperature reconstructions for individual years and decades are larger than those for longer time periods and because not all of the available proxies record temperature information on such short timescales.
This NAS statement provides fodder for both the believers and deniers. The deniers posit that if global temperature was indeed unusually warm from 950 to 1250 AD, at a time when CO 2 , CH 4 , and N 2 O were known to be stable ( Fig. 1.2), then other factors such as solar luminosity must be responsible. If so, the argument goes, then perhaps the late twentieth century and early twenty-first century warming is due to some factor other than anthropogenic GHGs.
A recent study of ocean heat content supports the view that the higher temperature of the MWP was indeed global in nature (Rosenthal et al. 2013), similar to the conclusion reached by Soon and Baliunas (2003) a decade earlier. If so, then the time series of ΔT shown in Fig. 1.2 is in need of revision. Regardless, as shown below, extremely strong scientific evidence implicates GHGs produced by human activities as the primary driver of rising global temperature during the past half-century.

The Anthropocene
The Anthropocene refers to the recent interval during which the atmospheric abundance of GHGs that drive Earth's climate have increased due to human activity (Crutzen and Stoermer 2000). Most peg the start of the Anthropocene to the mideighteenth century, linked to the invention of the steam engine by James Watt in 1784 (Steffen et al. 2015). Others suggest humans have had a discernable influence on GHGs for a much longer period of time, and argue for a start date to the Anthropocene as far back as 8000 years before present (ybp) (Ruddiman 2003).
We use 1765 as the start of the Anthropocene for several reasons. The largest influence of humans on GHGs has certainly occurred since 1765. Estimates of radiative forcing of climate (defined in Sect. 1.2.1) due to a wide variety of human activities based on a multi-year effort of scientists from many nations, are available in a transparent, easily accessible format 1 back to 1765 (Meinshausen et al. 2011).
Finally, changes in global mean surface temperature, to which RF of climate will be related, are much more certain from 1765 to present than for times extending back to the invention of agriculture (NAS 2006). Our choice is not meant to dismiss the importance of human influence on climate prior to 1765. If, as suggested by Ruddiman (2003), human activity 8000 ybp did indeed offset the onset of extensive glaciation due to declining summer insolation at Northern high latitudes (driven by Milankovitch orbital variations), this would be a fascinating benefit of human ingenuity, especially for indigenous peoples of high northerly latitudes.
Rather than wade deeper into the debate over the start of the Anthropocene, we next describe a few figures that illustrate the human fingerprint on the global carbon cycle and climate change over the past several centuries. Along the way, the mathematical principles needed to understand the material presented in Chaps. 2 and 3 are developed.

Radiative Forcing
In the absence of an atmosphere, the temperature of Earth would be governed by: Albedo, the Latin word for whiteness, refers to the fraction of incoming sunlight reflected to space (commonly about 0.3 (or 30 %) for Earth), S is the luminosity of our Sun at the distance of Earth's orbit (1370 W m −2 ), and σ is the Stefan Boltzmann constant (5.67 × 10 −8 W m −2 K −4 ). A value of S/4 is used because the Earth intercepts sunlight like a disk and radiates heat like a sphere; 1/4 is the ratio of the surface area of a disk to that of a sphere (this concept as well as Eq. 1.1 are explained in many introductory Earth Science textbooks). Solving for the putative temperature of planet Earth without an atmosphere yields 255 K, which is −18° Celsius (°C) or 0° Fahrenheit (°F). This is much colder than the average temperature of today's Earth. If the temperature found using Eq. 1.1 actually applied, our oceans would be frozen. The greenhouse effect, the trapping of radiation by our planet's atmosphere, is responsible for the difference between the Earth's actual temperature and that found using Eq. 1.1. Earth's mean surface temperature is about 15.5 °C or 60 °F. Earth's atmosphere is responsible for increasing the amount of energy the surface receives, by several hundred W m −2 , in comparison to an Earth devoid of an atmosphere. This excess heat is driven by the abundance and molecular properties of GHGs such as H 2 O, CO 2 , CH 4 , and N 2 O, as well as clouds (i.e., condensed H 2 O droplets).
Infrared radiation (or heat in the form of photons) emitted by Earth's surface is resonant with various vibrational modes of GHG molecules, inducing these photons to be absorbed and re-emitted in all directions. Some of this absorbed and re-emitted radiation is sent back to the surface. As such, GHGs in Earth's atmosphere act as a blanket, trapping heat that would otherwise escape to space. Water vapor, the most important GHG, is responsible for the majority of radiation sent back to the surface. The abundance of H 2 O in our atmosphere is controlled by thermodynamics: i.e., the evaporation of H 2 O from the oceans, condensation of H 2 O in the atmosphere to form clouds, and deposition of H 2 O back to the surface in the form of precipitation. Atmospheric H 2 O varied prior to the onset of human influence; the effect of thermodynamics on various isotopes of H 2 O preserved in ice cores is an important tool for quantitative reconstruction of past climate (Jouzel et al. 1987).
The radiative forcing (RF) of climate refers to the increase in the amount of heat directed to Earth's lower atmosphere as the abundance of GHGs rise. Here and throughout this book, we follow the convention established in the 2001 Intergovernmental Panel on Climate Change (IPCC) Physical Science Basic Report (IPCC 2001) that RF of climate is defined as the change in the net flow of energy (sunlight plus infrared heat) at the tropopause (boundary between the lower atmosphere, or troposphere, and the upper atmosphere, or stratosphere) relative to a particular start date, after allowing for stratospheric temperatures to adjust to radiative equilibrium. This concept is explained well in Sect. 2.2 of IPCC (2007).  (Meinshausen et al. 2011); (b) total anthropogenic ΔRF (red) and the global mean surface temperature anomaly (ΔT) relative to pre-industrial baseline (1850-1900 mean) from the modern instrument record (HadCRUT4) (Jones et al. 2012) and from various proxies (Jones and Mann 2004). See Methods for further information hence they are denoted ΔRF. 2 Red is used to represent warming (positive ΔRF); blue is used to show cooling (negative ΔRF). These ΔRF curves are based on the GHG and aerosol precursor abundances used to drive climate model simulations of IPCC (2013). The RF of climate due to human release of CO 2 , CH 4 , and N 2 O is the primary focus of this book. The historic ΔRF due to all anthropogenic GHGs (dashed red) exceeds that of the CO 2 /CH 4 /N 2 O triplet (dotted red) by a small amount, with most of the difference due to a class of compounds called Ozone Depleting Substances (ODS). Even though ODS exhibit a greenhouse effect, they are generally not labeled as GHGs because their most important detrimental effect is depletion of Earth's ozone layer. Also, industrial production of ODS has been successfully curtailed by the Montreal Protocol and the effect of these compounds on climate will diminish in the future (WMO 2014). The human release of pollutants that increase the burden of small particles in the troposphere, known as aerosols, leads to a reduction in the RF of climate (blue line, Fig. 1.3a). This occurs because many aerosols reflect sunlight. An estimate of ΔRF due to aerosols provided by Meinshausen et al. (2011) is shown in Fig. 1.3a. 3 As detailed below, there is considerable uncertainty in this term.
The total ΔRF due to human activity is shown by the solid red curve in Fig. 1.3a, b. All told, human activities have increased the RF of climate by about 2.3 W m −2 between 1750 and present. Figure 1.3b shows that the time variation of total ΔRF due to humans (red line) closely resembles the observed rise in global mean surface temperature anomaly (black and grey lines). Below we conduct quantitative analysis of these two quantities, both within this chapter as well as throughout Chap. 2. There are several aspects of Fig. 1.4 worth emphasizing. Human release of GHGs has warmed climate, with CO 2 being the most important contributor. The combined effect of the two other most important long-lived anthropogenic GHGs (CH 4 , N 2 O) plus all of the ODS compounds has enhanced this CO 2 -based warming by about 60 %. Tropospheric ozone (O 3 ), a pollutant harmful to human health and agriculture, has warmed climate over the course of the Anthropocene by nearly as much as CH 4 . Tropospheric O 3 is regulated by air quality regulations that vary by country and focus on surface conditions (i.e., the air we breathe). There is no coordinated international effort to limit future growth of tropospheric O 3 . Human release of pollutants that lead to formation of tropospheric aerosols causes climate to cool due to two processes nearly equal in magnitude but having large uncertainty: the reflection of sunlight by aerosols (Aerosol Direct Effect) and the effect of aerosols on cloud formation (Aerosol-Cloud Interaction). 5 The net effect of all human activity on ΔRF (bar labeled Total Anthropogenic) is about 25 % larger than ΔRF due to CO 2 and has a considerably larger uncertainty (size of respective error bars) (Chap. 8, IPCC (2013)).
Other aspects of global warming are represented in Fig. 1.4. The reflectivity of Earth's surface has changed primarily due to deforestation that makes the surface 5 The effect of anthropogenic aerosols on the radiative properties of clouds is different, from a climate modeling perspective, than the evolution of cloud properties as the surface warms. Specialists refer to the former as the aerosol indirect effect and the latter as cloud feedback. A considerable research effort informs us that the aerosol indirect effect leads to a cooling (negative RF) of uncertain magnitude and the cloud feedback could either lead to warming or cooling. brighter (leading to cooling) and the deposition of dark, carbonaceous material on snow that darkens the surface (leading to warming). Contrails from aircraft have led to a slight warming, mainly because of induced clouds. A slight warming term is also attributed to an increase in stratospheric humidity driven by rising tropospheric CH 4 , which is converted to H 2 O when lost in the stratosphere. The depletion of stratospheric O 3 has resulted in slight cooling. Finally and most importantly, scientists have been able to estimate the time variation of total solar irradiance over the course of the Anthropocene. The trend in solar irradiance over this two and a half century time period is small. Even if the MWP discussed above turns out to be as warm as the present decades, presumably due to an increase in solar irradiance during the middle of the Common Era (Bard et al. 2000), scientists are nonetheless confident the rise in temperature over the Anthropocene was driven by rising GHGs and not a change in the luminosity of our Sun (Chap. 8, IPCC (2013)).

Global Warming Potential
The global warming potential (GWP) metric was developed to guide public policy decisions regarding trade-offs of release of various GHGs. The GWP of a particular compound represents the ratio of the rise in global mean surface temperature (GMST) due to the release of a particular amount (mass) of this compound, relative to the rise in GMST resulting from the release of the same amount (mass) of CO 2 . Inherent in the computation of GWP is that the increase in GMST is found over a particular time horizon. The most commonly used time horizons are 20 and 100 years. The mathematical expression for the GWP of CH 4 over a 100-year time horizon, which relies on the use of a calculus concept called integration, is given by: where a CH 4 and a CO 2 are the radiative efficiencies (units W m −2 kg −1 ) of CH 4 and CO 2 , respectively, and CH 4 (t) and CO 2 (t) represent the time dependent response to the release into the atmosphere of the same mass of these two GHGs. 6 The atmospheric lifetime of a GHG denotes the time it takes for a pulsed release of the gas to decay by 1/e of the initial value, where e ≈ 2.718. The lifetimes of CH 4 6 Equation 1.2 represents a computer simulation of the cumulative radiative forcing of climate over a 100 year time period due to release of a pulse of a certain amount (mass) of CH 4 , divided by the cumulative radiative forcing over the same time period due to simulated release of the same amount of CO 2 . Since the pulse of CH 4 decays faster than the pulse of CO 2 , due to the shorter lifetime of CH 4 , the GWP of CH 4 is larger when shorter time periods are considered. and N 2 O used in IPCC (2013) calculation of GWP are 12.4 and 121 years, respectively. These lifetimes are typically used for evaluation of the numerator of Eq. 1.2, as most GWP estimates assume pure exponential decay. Conversely, the decay of the pulse of CO 2 in the denominator of Eq. 1.2 is found using a computer model of the global carbon cycle. As shown in Chap. 3, the GWP of various GHGs is vitally important for assessing the efficacy of the Paris Climate Agreement 7 . Table 1.1 provides the GWP of CH 4 and N 2 O from the past 4 IPCC reports. The GWP of GHGs has been updated over time due to evolving knowledge of the radiative efficiencies and lifetimes of atmospheric compounds. Also, IPCC (2013) provided two values of GWP for CH 4 and N 2 O: with and without consideration of carbon cycle feedback.
The most commonly used GWPs for public policy are those found for a 100-year time horizon. This preference is traceable to the 1997 Kyoto Protocol of the United Nations Framework Convention on Climate Change (UNFCCC), which was based on 100-year GWPs. Furthermore, since the values of GWP given in IPCC (2013) that do not allow for carbon cycle feedback are most analogous to values of GWPs given in prior IPCC reports, we will use GWPs for CH 4 and N 2 O of 28 and 265, respectively, in our analysis of the Paris Climate Agreement. The GWPs of other GHGs used in our analysis are based on Chap. 8 of IPCC (2013).
The fact that the GWPs of CH 4 and N 2 O are much larger than 1 means that, on a per mass basis, these GHGs are considerably more potent than CO 2 . 8 Furthermore, the GWP of CH 4 is much larger over a 20-year horizon than a 100-year time horizon, due to the 12.4 year lifetime for CH 4 used in the calculation of GWP. Integrated over 20 7 The Paris Climate Agreement was negotiated at the 21st Conference of the Parties (COP21) held in Paris, France during December 2015. The COP meetings are an annual gathering of representatives from participating nations, environmental agencies, and industry to address concerns of climate change. For more information see: http://unfccc.int/paris_agreement/items/9485.php 8 Some textbooks and reports provide GWP values on a per molecule basis, rather than a per mass basis. A molecule of CO 2 with atomic mass of 44 weighs 2.75 times a molecule of CH 4 (atomic mass of 16). Using the IPCC (2013) value for the GWP of CH 4 on a 100 year time horizon, without consideration of carbon cycle feedback, scientists would state CH 4 is 28 times more potent than CO 2 on a per mass basis and, at the same time, is 10.2 (28 ÷ 2.75) times more potent than CO 2 on a per molecule basis. years, a significant fraction of the initial, pulsed release of CH 4 is present in the modeled atmosphere. However, integrated over 100 years, a much smaller fraction is present. As discussed in Chap. 4, controlling inadvertent release of CH 4 to the atmosphere will likely be vitally important for reaching the goals of the Paris Climate Agreement, keeping warming well below 2.0 °C and aiming to limit warming to 1.5 °C. The importance of CH 4 would be amplified by a factor of 3 (ratio of 84/28 from Table 1.1) if climate change over a 20-year time horizon were used to guide public policy, placing even more stringent controls on the atmospheric release of this GHG. Carbon dioxide is the most important anthropogenic GHG for RF of climate (Figs. 1.3a and 1.4), despite the more potent nature of CH 4 and N 2 O, because human society has released to the atmosphere a much greater mass of CO 2 than other GHGs. Simply put, CO 2 is the greatest waste-product of modern society. We now turn our attention towards the human fingerprints on global warming as well as on the atmospheric build-up of CO 2 and other GHGs.

Human Fingerprints
As described in Sect. 1.1, Earth's climate has undergone vast variations on geologic time scales. Many of these climate shifts are directly tied to changes in atmospheric CO 2 . Studies of paleoclimate must also consider effects on global mean surface temperature (GMST) of continental plate alignment (Hay et al. 1990), the seasonal distribution of sunlight related to variations of Earth's orbit (Erb et al. 2013), as well as the radiative forcing of climate due to aerosols (Chylek and Lohmann 2008) and GHGs other than CO 2 (Sagan and Mullen 1972).
Even though Earth's climate and the abundance of GHGs co-vary on geologic time scales due to natural processes, a scientific consensus has nonetheless emerged that the recent rise in GMST as well as atmospheric burdens of CO 2 , CH 4 , and N 2 O are all driven by human activities (IPCC 2013). Here we briefly review some of the most important human fingerprints on temperature and GHGs. To place the material that follows in the proper perspective, it is important to understand that the time scales involved with geologic change and human history are enormously different. For instance, the ratio of the time since the rise of forests (400 Mypb) to the time since the advent of agriculture (~12,000 ybp) is enormous. If time on Earth since the rise of forests were compressed into a 24 h day, the time since the advent of agriculture would take 2.6 s, which is less than the time it takes to read this sentence!

Rising Temperature
As noted above, correlation does not demonstrate causation. We shall first examine the quantitative relation between the rise in the GMST anomaly (ΔT) shown in Fig. 1.3 and the change in radiative forcing of climate (ΔRF) attributed to humans over the Anthropocene.
The Stefan-Boltzmann equation relates the temperature of an object to the rate at which it is able to disperse energy: Power is used in Eq. 1.3 since Power is defined as Energy/Time, and the Stefan-Boltzmann equation is based on the rate at which energy is dispersed; σ is the same constant used in Eq. 1.1. In equilibrium, Earth's surface releases (or radiates) energy at the same rate it is supplied to the surface by the atmosphere. Hence Power in Eq. 1.3 can be replaced with RF, where RF represents the atmospheric radiative forcing of climate: Those who have taken calculus will understand that upon taking the derivative of Eq. 1.4 and re-arranging terms, it can be shown that: where ΔRF represents a perturbation to the system (i.e., the rise in RF of climate due to human release of GHGs) and ΔT represents the response (i.e., resulting rise in global mean surface temperature).
Scientists use relations such as Eq. 1.5 to diagnose the output of climate models.
A common value for the term 1 4 3 sT in Eq. 1.5 is 0.31 K/(W m −2 ), which is related to the temperature at which Earth radiates to space (Bony et al. 2006). Substituting this numerical value for 1 4 3 sT into Eq. 1.5 leads to: We now examine the quantitative consistency between the rise in temperature (ΔT in Fig. 1.3) over the Anthropocene and the radiative forcing of climate attributed to humans (ΔRF in Fig. 1.3). The product 0 31 23 is equal to 0.7 K (which is the same as 0.7 °C), 9 quite close to the observed rise in GMST (about 0.9 °C) over the course of the Anthropocene. As examined in detail in Chap. 2, the actual relationship between ΔT and ΔRF requires a consideration of factors such as climate feedback (i.e., enhancement or diminution of the RF of climate imposed by humans due to changes in factors such as atmospheric humidity and clouds) as well as the transport of heat from the atmosphere to the world's oceans, which specialists refer to as ocean heat export. It is likely, for instance, that positive feedback (enhancement) of the direct RF of climate caused by humans is responsible for the difference between the expected (0.7 °C) and observed (0.9 °C) rise in GMST over the course of the Anthropocene. Most importantly, the calculation conducted above reveals quantitative consistency between the observed and expected rise in global temperature over the course of the Anthropocene. This is a critically important first step in the attribution of global warming to humans.
Several other aspects of global warming bear the human fingerprint. Climate models predict that as GHGs rise, the lowest layer of the atmosphere (the troposphere) will warm while the second layer of the atmosphere (the stratosphere) should cool. Stratospheric cooling is a consequence of the blanketing effect of GHGs: as atmospheric levels of GHGs rise, a larger fraction of the thermal energy emitted by the surface is absorbed, re-emitted, and therefore blocked from reaching the stratosphere. As shown in Fig. 1.5, tropospheric warming coupled with stratospheric cooling is seen in the climate record, at least over the part of the Anthropocene for which modern measurements of atmospheric temperature profiles exist (Sherwood and Nishant 2015). About two-thirds of the cooling of the upper stratosphere for the time period 1979-2005 has been attributed to rising GHGs, with the remainder attributed to humaninduced depletion of stratospheric O 3 (Mitchell 2016). The pattern of temperature changes with respect to altitude and latitude throughout the troposphere and stratosphere agrees with the pattern predicted by climate models to a high degree of statistical significance (Santer et al. 2013a), although these models do tend to overestimate the amount of warming observed in the lower atmosphere (Santer et al. 2013b). 10 Another important human fingerprint of global warming is the observation that the altitude of the tropopause, the boundary between the troposphere and the stratosphere, has been rising as predicted by climate models (Santer et al. 2013a). Had modern global warming been due to an increase in the luminosity of the Sun or a release of energy from the world's oceans, scientists would have expected to observe warming in the stratosphere and troposphere as well as little to no change in the height of the tropopause.

Carbon Dioxide
Carbon dioxide (CO 2 ) is the single greatest waste product of modern society. There is compelling scientific evidence that the rise in atmospheric CO 2 during the Anthropocene is due, nearly entirely, to human activity. The rise in CO 2 from 1765 to the early 1900s was predominately driven by the clearing of forests for agriculture (also known as land use change, or LUC) (Siegenthaler and Oeschger 1987). For a few decades subsequent to 1900, LUC and the combustion of fossil fuels made nearly equal contributions to the rise in atmospheric CO 2 . Since the early 1950s, the growth of atmospheric CO 2 has been driven primarily by the combustion of fossil fuels 11 (Le Quéré et al. 2015).
Precise quantification of the contemporary abundance of atmospheric CO 2 was initiated by Charles David Keeling at the Mauna Loa Observatory (MLO) in Hawaii during March 1958, as part of the International Geophysical Year (IGY) program (Keeling et al. 1976). On the first day of measurement, an atmospheric CO 2 abundance of 313 parts per million (ppm) was recorded. 12 The MLO CO 2 record is a signature accomplishment of the IGY, carried out from July 1957 to December 1958 and characterized by international cooperation on many scientific fronts. The ability of nations such as the United States and the Soviet Union to collaborate on IGY, despite the Cold War, serves as an inspiration for the level of international cooperation that will be needed to address the consequences of rising GHGs recorded at the Mauna Loa observatory.
On the day this sentence was written, atmospheric CO 2 at MLO was 407.66 ppm. This CO 2 reading amounts to a thirty percent increase relative to Keeling's first 12 Mixing ratio denotes the fraction of all air molecules that exist as a particular compound. Keeling's initial observation means 313 out of every million air molecules were present as CO 2 . The history of the Mauna Loa Observatory as well as an account of this initial observation are at https://scripps.ucsd.edu/programs/keelingcurve/2013/04/03/the-history-of-the-keeling-curve  (2015) except we use altitude, rather than pressure, for the vertical coordinate. See Methods for further information observation and a forty-five percent increase relative to 280 ppm, the atmospheric mixing ratio of CO 2 commonly assumed to have been present at the start of the Anthropocene. Daily measurements of atmospheric CO 2 are provided at various websites, including https://www.co2.earth/daily-co2. We now describe the scientific evidence that humans are responsible for the rise of CO 2 . Our focus is on 1959 to present, the modern instrument era. Readers interested in learning about the human impact on CO 2 over the earlier part of the Anthropocene (i.e., prior to 1959) are encouraged to examine studies such as Siegenthaler and Oeschger (1987), Ruddiman (2003), Le Quéré et al. (2015), and Steffen et al. (2015). Figure 1.6 shows time series of the atmospheric build-up of CO 2 and fossil fuel emissions of CO 2 , from 1959 to present. Measurement of CO 2 at MLO (Keeling et al. 1976) and an estimate of global mean CO 2 provided by the US National Oceanographic and Atmospheric Administration (NOAA) Earth System Research Laboratory (ESRL) (Ballantyne et al. 2012) are shown in Fig. 1.6a. The saw-tooth pattern of the MLO CO 2 reveals the breathing of the biosphere: seasonal minimum occurs in late boreal summer just before deciduous trees, which predominantly exist in the NH, begin to drop their leaves. Seasonal maximum occurs in mid-spring of the NH, just before trees and plants bloom. The global, annual record of CO 2 exhibits a steady upward march over the past six decades. Figure 1.6b provides our first evidence that humans are responsible for the rise of CO 2 over the past six decades. This panel compares the annual, global release of CO 2 to the atmosphere due to the combustion of fossil fuels (Boden et al. 2013) and land use change (Houghton et al. 2012) (green bars) to the annual rise in global atmospheric CO 2 (blue bars); both quantities are expressed in units of 10 9 metric tons of CO 2 (Gt CO 2 ) emitted per year. 13 In some years, such as 1977, 1979, 1987, 1988, and 1998, the rise in atmospheric CO 2 is more than half of the CO 2 input to the atmosphere by humans (i.e., the height of the blue bar is more than half the height of the green bar). Typically, the annual rise in the mass of atmospheric CO 2 (blue bars) equals between 40 and 50 % of the mass of CO 2 released to the atmosphere by humans (green bars). This comparison demonstrates quantitative plausibility that the observed rise in atmospheric CO 2 during the modern instrument era was indeed due to human activity. Figure 1.7 illustrates the three most important pieces of observational evidence that scientists use to reveal the human fingerprint on rising CO 2 . Time series of the mixing ratio of atmospheric CO 2 measured at MLO in Hawaii (19.82°N latitude) are compared to CO 2 measured at the South Pole Observatory (SPO) in Fig. 1.7a. Figure 1.8 compares the difference between annual averages of CO 2 at MLO and SPO (ΔCO 2 MLO−SPO ) for specific years plotted against the total human release of atmospheric CO 2 for each particular year. Figures 1.7a and 1.8 show that CO 2 is higher in the NH than the Southern Hemisphere (SH). This hemispheric gradient has long been used as evidence for the human influence on atmospheric CO 2 , since anthropogenic emissions occur predominantly in the NH (Tans et al. 1990). The strong correlation of ΔCO 2 MLO−SPO versus total human release of CO 2 shown in Fig. 1.8 further demonstrates that the anthropogenic activity exerts primary control on the hemispheric gradient in atmospheric CO 2 (Fan et al. 1999). Figure 1.7b illustrates the small decline of atmospheric O 2 , which is another fingerprint of the human influence on rising CO 2 (see Methods for an explanation of the units and numbers). As CO 2 is released to the atmosphere by the combustion of fossil fuel, the oxygen (O) content of the newly emitted CO 2 comes from atmospheric molecular oxygen (O 2 ). 14 If rising atmospheric CO 2 were due to volcanic 14 Fossil fuels are characterized by a mixture of hydrogen (H), carbon (C), and depending on the source other elements such as oxygen (O), nitrogen (N), and sulfur (S). Chemical compositions range from CH 4 (methane or natural gas), C 8 H 18 (octane, commonly used to represent automotive gasoline), to C 135 H 96 O 9 NS (coal, which can also contain other elements such as arsenic, lead, mercury, etc.). Since H and C are the dominant elements, fossil fuels are commonly called hydrocarbons. The O in CO 2 produced by combustion of methane or gasoline originates entirely from (c) monthly, Tropical Pacific ENSO Index (Zhang et al. 1997): periods of dark red longer than 5 months indicate an El Niño event; (d) ratio of ocean plus biospheric uptake of CO 2 divided by total emissions of CO 2 (i.e., difference between the height of the green bar and blue bars in panel (b) divided by the height of the green bar) (grey) as well as 3 year running mean (black) and trend-line of a linear least squares fit to the 3 year running mean (blue). See Methods for further information activity or deep sea vents, atmospheric O 2 would be unaffected because the dominant form of outgassed carbon is CO 2 , with the O drawn from abundant oxygen in Earth's crust. Measurement of atmospheric O 2 with sufficient precision to quantify the minute, putative decline was an instrumental challenge first overcome by Ralph Keeling (Keeling et al. 1996), the son of Charles David Keeling. The slight decline in atmospheric O 2 recorded in Fig. 1.7b provides strong quantitative evidence that combustion of fossil fuel is the driving factor behind rising CO 2 . The final human fingerprint involves the isotopic composition of atmospheric CO 2 . The most common form of carbon has an atomic mass of 12 ( 12 C), due to the presence of six protons and six neutrons in the nucleus. 15 Carbon can exist in two other forms: carbon 13 ( 13 C) and carbon 14 ( 14 C), due to the presence of either 7 ( 13 C) or 8 ( 14 C) neutrons in the nucleus. The chemical properties of a compound are determined by the number of electrons, which equals the number of protons if a compound is neutral atmospheric O 2 , whereas the O in CO 2 produced by combustion of coal originates mainly from atmospheric O 2 . 15 Atomic mass is the sum of the number of protons and neutrons. (i.e., uncharged). Hence 12 C, 13 C, and 14 C are all considered to be different forms of carbon because they all contain six electrons. Most importantly, biological properties of a compound are mass dependent: our bodies prefer 12 C over the other two heavier forms of carbon that are digested, because lighter molecules diffuse more readily through our capillaries. The term isotopic composition, as used here, refers to the relative abundance of 13 C in a sample of atmospheric CO 2 compared to the sum of 12 C, 13 C, and 14 C in the same sample, and is expressed using the notation δ 13 C. Figure 1.7c shows a time series of δ 13 C recorded at MLO (Keeling et al. 2005). The downward decline of δ 13 C means atmospheric CO 2 is getting isotopically lighter over time. In other words, at the start of the time series in 1980, the relative proportion of 13 C to 12 C in atmospheric CO 2 at Hawaii was larger than today. This serves as our final fingerprint because the carbon content of fossil fuels, which formed from the decomposition of plants on geologic time-scales, are isotopically light relative to contemporary atmospheric CO 2 (Whiticar 1996). If rising levels of atmospheric CO 2 during the time period shown in Fig. 1.7c had been due primarily to volcanoes, atmospheric CO 2 would have been expected to have gotten isotopically heavier (Rizzo et al. 2014), which is the opposite of what has been observed.
It is stated in Sect. 1.1 that over geologic time scales, atmospheric CO 2 is controlled by volcanic activity and deep sea vents. Yet CO 2 shows no volcanic influence over the time of the modern instrument record. To further illustrate the lack of recent volcanic influence, the orange triangles in Fig. 1.6b have been placed at the time of eruption of Mount Agung, El Chichón, and Mount Pinatubo, the three largest eruptions over the past six decades. The growth of atmospheric CO 2 during the years of these eruptions (1963, 1982, and 1991) is unremarkable compared to other years: in fact, the growth of  (Keeling et al. 1976) minus CO 2 at the South Pole Observatory (SPO) (Tans et al. 1990) (vertical axis) versus annual total emission of CO 2 to the atmosphere from the combustion of fossil fuels (Boden et al. 2013) and land use change (Houghton et al. 2012). Data span the time period 1959-2015. Numerical results of a linear least squares fit as well as the correlation coefficient are also given. See Methods for further information atmospheric CO 2 (blue bars) after the eruption of Mt. Agung (1963) and Mt. Pinatubo (1991) is suppressed relative to prior years. The best estimate of contemporary release of atmospheric CO 2 by volcanoes and deep sea vents reveals release of about 0.26 Gt of atmospheric CO 2 per year (Marty and Tolstikhin 1998), less than 1 % of the human burden. Interestingly and with touching irony for those who refuse to accept the human influence on global warming, the volcanic release of CO 2 during the 9 h explosive phase of Mt. Pinatubo on 15 June 1991 likely matched the total, global release of CO 2 by humans on that day (Gerlach 2011). More than 20,000 days have passed since the start of modern measurements of atmospheric CO 2 . On one day and one day only, global human release of atmospheric CO 2 was likely matched by a volcano. Human release of CO 2 has dwarfed volcanic release on the other 19,999 days.
Why have volcanoes been so dominant in the past, yet so unimportant in the present? One factor is that modern human civilization has not yet experienced a volcanic eruption of the magnitude known to have occurred in the past. The Volcanic Explosivity Index (VEI) denotes the size of volcanic eruptions (Newhall and Self 1982), much like the Richter Scale for earthquakes. Mt. Agung, El Chichón, and Mt. Pinatubo had VEIs of 5, 5, and 6, respectively. The most violent eruption Earth has experienced over the past 36 million years was the VEI of 9.1-9.2 eruption of La Garita 16 about 27.8 Mypb (Mason et al. 2004). Since the VEI scale is logarithmic with respect to volume of ejecta, a VEI 9 eruption would eject about 1000 times more mass than Mt. Pinatubo. 17 Had Mt. Pinatubo been VEI 9, it may have matched human emission of CO 2 over the prior 1000 days. In this case, of course, the ejection of CO 2 by such a monstrous event would have been the least of our concerns.
The other factor responsible for the minor role of volcanoes with respect to contemporary atmospheric CO 2 is that Earth is presently in a geologically dormant period. The Deccan Traps of India is one of the largest, most well-studied ancient volcanic features on Earth. Eruptions of this massive province, approximately 65 Mybp, may have been characterized by a decades-long explosive events (Self et al. 2006). The perturbation to atmospheric CO 2 by the Deccan Traps is the subject of active research, with some studies (Dessert et al. 2001) suggesting a considerably larger influence than others (Self et al. 2006).
We conclude by providing a brief overview of the latest understanding of the factors that control atmospheric CO 2 . More detailed information, updated annually, is maintained at http://www.globalcarbonproject.org/carbonbudget.
It is well established that a substantial portion of the CO 2 released to the atmosphere by human activity is absorbed by trees and plants (i.e., the terrestrial biosphere) as well as the world's oceans (Le Quéré et al. 2015). Uptake of anthropogenic CO 2 by plants is facilitated by three factors: higher levels of atmospheric CO 2 promote faster growth of plants (Zhu et al. 2016), global warming has increased the length of the growing season (Le Quéré et al. 2015), and human supply of fixed nitrogen to the biosphere promotes a more fertile environment for plant growth (Galloway et al. 2014). The world's oceans contain a mass of carbon about 50 times greater than that in the atmosphere and it has long been known that the world's oceans would uptake a portion of the CO 2 placed into the atmosphere by human activities (Revelle and Suess 1957).
An atmospheric and oceanic phenomena known as El Niño Southern Oscillation (ENSO) has been extensively studied as a driving factor for the variation of the height of the blue bars (atmospheric growth of CO 2 ) relative to the green bars (human release of CO 2 ) shown in Fig. 1.6b (Keeling et al. 2005;Randerson et al. 2005;Zeng et al. 2005). When the index shown in Fig. 1.6c is shaded dark red for a period of ~5 months or longer, the tropical ocean/atmosphere system is in the midst of an ENSO event. 18 The growth of atmospheric CO 2 tends to be larger than normal for about a year after the peak of an ENSO event, with the effect maximizing about 6 months after the peak (Zeng et al. 2005). An ENSO event affects atmospheric CO 2 due to suppression of oceanic uptake as well as the tendency for human-set fires to occur in drought stricken regions during certain ENSO years (Randerson et al. 2005). During late 2015, Earth experienced another major ENSO event, which likely was responsible for the more rapid rise of atmospheric CO 2 in 2015 compared to prior years. Indeed, the preliminary estimate of total human release of CO 2 in year 2015 given by Le Quéré et al. (2015), which is the origin of the last green bar in Fig. 1.6b, shows a slight decline relative to 2014. Should this decline in human release of CO 2 continue in future years, the height of the blue bars in future updates to Fig. 1.6b will fall relative to the value for 2015, except for years marked by either large ENSO events and/or extensive biomass burning.
The fraction of anthropogenic CO 2 removed each year via the world's terrestrial biosphere and oceans is depicted by the grey bars in Fig. 1.6d. There is considerable year-to-year variability, which has been widely studied and is attributed mainly to terrestrial biosphere (Bousquet et al. 2000;Le Quéré et al. 2003). Averaged over the entire data record, 56 % of the CO 2 released to the atmosphere by humans by the combustion of fossil fuels and land use change has been absorbed by land and ocean sinks. In other words, the actual rise in atmospheric CO 2 equals about 44 % of that known to have been emitted by humans.
The efficiency of the combined land and ocean sink for atmospheric CO 2 appears to be weakening over time. Figure 1.6d contains two lines. One shows a 3 year running mean (black) of the numerical values of each grey bar, for data starting in 1959 and ending in 2014. Values for 2015 are excluded from the 3 year running mean, because data for this year are considered preliminary at the time of writing. An entity such as a 3 year running mean is a common statistical method used to analyze data that exhibit a large amount of year-to-year variability, such as the grey bars in Fig. 1.6d. The trend-line (blue) shows a linear least squares fit to the 3 year running mean, another common technique used to examine geophysical data. The trend-line has a slope of −0.0013 per year, which means the fraction of anthropogenic CO 2 removed by the combined land and ocean sink may have declined from about 0.6 in 1959 to about 0.53 in 2014. However, there is considerable uncertainty (in this case, 18 During an ENSO event warm waters in the Tropical Western Pacific ocean migrate to the Central and Eastern Pacific, causing shifts in the location of oceanic upwelling and atmospheric storms, as well as significant perturbations to the global carbon cycle. An informative animation of ENSO is provided at http://esminfo.prenhall.com/science/geoanimations/animations/26_NinoNina.html ±0.0014 per year) in the slope of the trend fit line. This uncertainty encompasses (albeit, just barely) the possibility that the combined land and ocean sink may not have actually changed and also, at the same time, another possibility that this uptake could have changed from 0.6 in 1959 to 0.45 in 2014. This analysis builds upon the work and supports the conclusions of Le Quéré (2010), who also emphasize the urgent need to reduce the uncertainty in the time rate of change of the combined land and ocean sink for human release of atmospheric CO 2 . If the efficiency of the combined land and ocean sink for CO 2 is truly declining over time, then this is enormously important for the response of society to anthropogenic release of GHGs.

Methane
Methane (CH 4 ) is a vitally important anthropogenic GHG. The atmospheric abundance of CH 4 has risen from a pre-Anthropocene value of 0.7 ppm to a contemporary abundance of 1.84 ppm ( Fig. 1.2). The rise in CH 4 between 1750 and 2011 has induced a RF of climate of 0.48 W m −2 ( Fig. 1.4), second only to the RF of CO 2 among anthropogenic GHGs. 19 Methane is therefore commonly referred to as the second most important anthropogenic GHG.
Studies of atmospheric CH 4 are numerous, complex, and quite varied, owing to a variety of natural and human sources (see Kirschke et al. (2013) and references therein). Figure 1.9 shows an estimate of the sources (i.e., flux into the atmosphere) of CH 4 , in units of 10 12 g of CH 4 (Tg CH 4 ) emitted per year, 20 averaged over the decade 2000-2009 from Conrad (2009) and Kirschke et al. (2013). The figure also contains an estimate of the sinks (i.e., atmospheric loss) of CH 4 over the same period of time.
A number of scientifically important details regarding atmospheric CH 4 are contained in Fig. 1.9. First, the magnitude of the source is slightly larger than the sink, consistent with the fact that atmospheric CH 4 is rising. Also, there are various human and natural sources of considerable magnitude. As noted above, wetlands are the largest natural source of CH 4 . Other natural sources include termites and the release of CH 4 from gas hydrates. 21 Finally, anthropogenic production of CH 4 occurs due to many aspects of our industrialized world, including the fossil fuel industry, 19 The RF of climate due to CO 2 over the same time period was 1.82 W m −2 . The notion that CH 4 is a more potent GHG than CO 2 is reconciled with these two RF estimates upon realization that the rise of the atmospheric mixing ratio of CO 2 over the Anthropocene, 120 ppm, is about 106 times the rise of CH 4 . For those who would like to dig into the numbers, radiative efficiencies of CO 2 and CH 4 are needed. In mixing ratio units, these radiative efficiencies are 1.4 × 10 −2 W m −2 per ppm for CO 2 and 3.7 × 10 −1 W m −2 per ppm for CH 4 (see Table TS.2 of IPCC (2007)). A "back of the envelope" estimate for the expected RF due to CH 4 is then: [1.82 W m −2 × (3.7 × 10 −1 ÷ 1.4 × 10 −2 )] ÷ 106 = 0.45 W m −2 . This estimate for the RF of CH 4 over the Anthropocene is quite close to the actual IPCC (2013) value of 0.48 W m −2 , which was found in a much more computationally intensive manner. 20 Tera is derived from the Greek word teras, meaning monster, and is often used as a prefix to denote 10 12 , or a trillion. A mass of 1 Tg (10 12 g) is the same as one thousandth of a giga tonne, where tonne refers to metric ton. 21 Methane hydrates are water ice structures that contains gaseous CH 4 in the core, and are prevalent in continental margins (Kvenvolden 1993). human set fires (biomass burning), microbial processes in the stomachs of ruminants, 22 as well as anaerobic conditions common in rice paddies and landfills. As detailed in Chap. 4, if human release of methane is to be curtailed, many aspects of modern society will need to be addressed, including how we heat our homes, generate our electricity, and produce our food.
Two considerable complications for the proper accounting of the human release of CH 4 are posed by possible alteration of the wetland source due to climate-change induced changes of the hydrologic cycle (i.e., floods and drought) as well as the possible release of prodigious amounts of CH 4 from the Arctic as permafrost thaws, due to global warming (Koven et al. 2011). For now, at least, the source of CH 4 due to Arctic permafrost is small on a global scale (Kirschke et al. 2013).
We turn our attention to the scientific importance of the numerical estimates of the CH 4 source and sink strengths shown in Fig. 1.9. The globally averaged sink for CH 4 is 550 Tg per year. The mass of CH 4 in the atmosphere, at present, 23 is about 5326 Tg. The atmospheric lifetime of CH 4 is given by: 22 Ruminants are mammals such as cattle, sheep, deer, giraffes, etc. that acquire nutrients by fermenting plant-based foods in a specialized stomach prior to digestion. 23 We can approximate the mass of CH 4 in the atmosphere by multiplying the mass of the entire atmosphere, 5.2 × 10 21 g, by the mixing ratio of CH 4 , which is 1.84 ppm or 1.84 out of every million air molecules. We must also account for the ratio of the atomic mass of CH 4 (16) to the mean atomic mass of air (28.8). The atmospheric mass of CH 4 therefore equals 1.84 × 10 −6 × 5.2 × 10 21 g × (16/28.8) = 5.326 × 10 15 g = 5326 Tg. (1.7) On average, a molecule of CH 4 released to the atmosphere will persist for about a decade until it is removed by either a chemical reaction or a soil microbe. Prather et al. (2012) report a present-day CH 4 lifetime of 9.1 ± 0.9 years, consistent with our calculation above. Conversely, a lifetime of 12.4 years for CH 4 was used by IPCC (2013) in the calculation of GWPs because it is thought that the release of a large pulse of CH 4 to the atmosphere will prolong the atmospheric lifetime, due to resulting changes in the chemical composition of the tropical troposphere.
The ~10 year atmospheric lifetime for CH 4 has important policy implications. This is best illustrated by comparing the human release of CH 4 to that of CO 2 . Throughout the world, humans presently release about 335 Tg of CH 4 and 39 Gt of CO 2 per year. Since 1000 Tg = 1 Gt, these sources are 0.335 Gt of CH 4 and 39 Gt of CO 2 per year: i.e., the mass of CO 2 released to the atmosphere each year by human society is about 116 times more than the mass of CH 4 . The impact on climate is entirely dependent on the time scale of interest. Nearly all of the CH 4 released to the atmosphere in year 2015 will be gone by the end of this century. The CO 2 -equivalent emission of CH 4 , found by multiplying the current release by the GWP for CH 4 for a 100-year time horizon, is 28 × 0.335 Gt of CH 4 or 9.4 Gt per year. If our concern is global warming over the next century, then we would conclude the human release of CO 2 in year 2015 was about four times more harmful for climate (39 ÷ 9.4 = 4.1) than the release of CH 4 . However, if our concern is the next two decades, we must consider the GWP of CH 4 over a 20-year time horizon. In this case, the CO 2equivalent emission of CH 4 is 84 × 0.335 Gt or 28.1 Gt per year, and we would conclude the present human release of CH 4 is nearly as harmful for climate (28.1 versus 39) as the release of CO 2 .
As noted above, international policy for the regulation of GHGs generally utilizes GWPs found over a 100-year time horizon. Perhaps this is appropriate, given CO 2 is such a long-lived GHG (i.e., a CO 2 molecule released today by humans will likely persist in the atmosphere longer than a molecule CH 4 ). However, should the world ever face an impending climate catastrophe in the midst of rapidly rising abundances of both atmospheric CO 2 and CH 4 , the greatest leverage for nearimmediate relief will be to reduce anthropogenic emissions of CH 4 ) or other short-lived pollutants (Pierrehumbert 2014). Of course this is much easier stated than accomplished given the wide variety of human activities that release CH 4 , as well as the tendency for energy production in the United States to become increasingly more CH 4 -based, given the abundant source of natural gas now being extracted by fracking. 24 Considerable research effort has been directed towards quantification of the anthropogenic versus natural sources of atmospheric CH 4 due to the scientific importance of this apportionment. Suppose, as is likely, that CH 4 had the same (or nearly the same) lifetime for removal from the atmosphere as today, for conditions that prevailed prior to the Anthropocene. Also, if the natural source of atmospheric CH 4 was the same (or similar) for these two time periods, it can be shown that: where Source Natural is the present natural flux of CH 4 to the atmosphere and Source Natural+Human is the total flux. Using the numerical values for these two fluxes 25 from Fig. 1.9, which are based on Table 1 of Kirschke et al. (2013), yields an estimate for CH 4 Pre-Anthropocene /CH 4 Present of 0.39. This estimate is astonishingly close to the actual ratio of CH 4 Pre-Anthropocene /CH 4 Present = 0.38, found using the atmospheric abundances given in the opening paragraph of this section. 26 Thus, an analysis of the sources of atmospheric CH 4 for the contemporary atmosphere provides strong quantitative support for the notion that human activities are indeed responsible for the rise of atmospheric CH 4 over the course of the Anthropocene.

Nitrous Oxide, Ozone, and Ozone Depleting Substances
Nitrous oxide (N 2 O) is commonly considered to be the third most important anthropogenic GHG. The atmospheric abundance of N 2 O has risen from a pre-Anthropocene value of 0.273 ppm to a contemporary abundance of 0.329 ppm ( Fig. 1.2). The rise in N 2 O between 1750 and 2011 has induced a RF of climate of 0.17 W m −2 ( Fig. 1.4).
Nitrous oxide is long-lived, with a lifetime of about 120 years. The vast majority of the atmospheric loss of N 2 O occurs in the stratosphere (Minschwaner et al. 1993). Nitrous oxide has a GWP of 264 (20-year time horizon) or 265 (100-year time horizon) (Table 1.1) according to IPCC (2013). The GWP of N 2 O is nearly the same for both time horizons because a pulse of N 2 O released to the atmosphere decays, within models used to calculate GWPs, in a manner quite similar to the decay of a pulse of CO 2 .
Current best understanding of the human sources of N 2 O is described in Chap. 6 of IPCC (2013). The total anthropogenic source is estimated to be 21.7 Tg of N 2 O per year, 27 albeit with considerable uncertainty. The human source could be as low 25 Source Natural = 218 Tg year −1 (total of the human terms; i.e., height of the six rectangles to the right of "Human" in Fig 1.9); Source Natural+Human = 553 Tg year −1 (total of all sources, Fig. 1.9). 26 These abundances yield CH 4 Pre-Anthropocene /CH 4 Present = 0.7 ppm/1.84 ppm = 0.38. 27 The IPCC (2013) best estimate for human release of N 2 O is 6.9 Tg of nitrogen per year, but we must convert to N 2 O to make use of the GWP of N 2 O. 6.9 Tg of N per year is the same as 6.9 × (44 ÷ 14) = 21.7 Tg of N 2 O per year, where 44 and 14 are the atomic masses of N 2 O and N. as 8.5 or as high as 34.9 Tg N 2 O per year according to Table 6.9 of IPCC (2013). Prather et al. (2012) report a smaller best estimate for the human source of 20.4 Tg N 2 O per year, with a reduced uncertainty of ±4 Tg N 2 O per year. Agriculture is the dominant activity responsible for human release of N 2 O: use of nitrogen fertilizers results in release of N 2 O to the atmosphere due to microbial processes in soils (Smith et al. 1997). Contemporary human emissions of N 2 O presently make a contribution to global warming 28 that is ~15 % that of emissions of CO 2 .
The largest natural sources of N 2 O are production from soils that lie beneath vegetation unperturbed by humans and release from the world's oceans. The natural source is estimated to be 34.6 Tg of N 2 O per year, again with considerable uncertainty (range from 17.0 to 61.6 Tg of N 2 O per year) (IPCC 2013). Large uncertainties for both the human and natural sources of N 2 O, as well as the long atmospheric lifetime for N 2 O, preclude meaningful use of Eq. 1.8 to examine the consistency between the rise in N 2 O and our understanding of the natural and anthropogenic source strengths. Nonetheless, the long-term rise in N 2 O since 1977, the observation of larger abundances in the NH than the SH documented on websites such as http:// www.esrl.noaa.gov/gmd/hats/combined/N2O, and field measurements of strong anthropogenic sources (Table 6.9 of IPCC (2013)) all provide strong scientific evidence that humans are responsible for the vast majority of the rise in N 2 O over the course of the Anthropocene.
The possible increase in atmospheric N 2 O due to expanded use of biofuels will receive considerable attention in the next few decades. There is considerable interest in the development of biofuels as a replacement for fossil fuels because, in theory, biofuels could be close to carbon neutral. The notion of carbon neutrality is predicated on the fact that the carbon in a hydrocarbon fuel produced by recent photosynthesis has been drawn out of the atmosphere just prior to combustion: i.e., the carbon is recycled. One of the many concerns regarding the modern biofuel industry is that the associated increase in production of atmospheric N 2 O due to the need for additional fertilizer will offset the climate benefit from the supposed carbon neutrality of this new fuel source (Crutzen et al. 2016).
The effect of N 2 O on stratospheric O 3 will also likely receive attention by researchers. Loss of N 2 O occurs in the stratosphere and, upon decomposition, N 2 O produces compounds that deplete stratospheric ozone (Ravishankara et al. 2009). Most interestingly, the ozone depletion potential 29 of N 2 O depends on future atmospheric abundances of CO 2 and CH 4 (Revell et al. 2015). Not only are CO 2 , CH 4 , and N 2 O (as well as chlorofluorocarbons, or CFCs) all important for climate, but these compounds are also inextricably linked for the future recovery of Earth's ozone layer.
While it is certainly true that most scientists consider N 2 O to be the third most important anthropogenic GHG, it is worth noting that the contribution to the RF of climate over the course of the Anthropocene by N 2 O is smaller than that of both Ozone Depleting Substances (ODS) and tropospheric O 3 (Fig. 1.4). Why then is N 2 O commonly considered to be the third most important anthropogenic GHG? The answer is nuanced but provides insight into the multi-disciplinary nature of modern atmospheric science.
The category labeled ODS in Fig. 1.4 consists of many gases (see Methods), including numerous CFCs, hydrochlorofluorocarbons (HCFCs), carbon tetrachloride (CCl 4 ), methyl chloroform (CH 3 CCl 3 ), etc. Industrial production of this class of compounds has been successfully regulated by the Montreal Protocol and subsequent amendments due to the harmful effects of these chemicals on Earth's protective ozone layer (WMO 2014). It is not commonly appreciated, but the climate protection accomplished by the Montreal Protocol (due to reduction in the atmospheric abundance of ODS that would have otherwise occurred) far exceeds the climate protection accomplished by the Kyoto Protocol (Velders et al. 2007). In other words, the positive RF of climate due to ODS in Fig. 1.4 would have been much larger had industrial production of these compounds not been halted by the Montreal Protocol. Nonetheless, most scientists do not apply the GHG label to the class of chemical compounds that deplete Earth's ozone layer. Also, none of the ODS compounds, alone, has a RF of climate as large as N 2 O. So N 2 O survives this challenge to its third place status.
The category labeled tropospheric O 3 in Fig. 1.4 also exerts a RF of climate that exceeds that due to N 2 O. Over the course of the Anthropocene, human release of chemicals such as carbon monoxide and nitrogen oxides produced by biomass burning and the combustion of fossil fuels has led to a build-up of tropospheric ozone, exerting a considerable influence on the RF of climate (Fig. 1.4). There has also been a slight cooling effect to the decline in stratospheric O 3 over the Anthropocene. Lack of consideration of tropospheric O 3 as the third most important anthropogenic GHG is due to various factors, including: (a) tropospheric O 3 is not emitted directly by humans but rather is produced in the atmosphere following chemical reactions of O 3 precursors released by humans; (b) surface O 3 , which is an important subcategory of tropospheric O 3 , is regulated by air quality agencies throughout the world (i.e., O 3 poses more harm to air quality than to climate); (c) all of the other anthropogenic GHGs tend to be long lived (atmospheric lifetimes greater than a year) and have nearly uniform global distributions, whereas tropospheric O 3 is short lived (atmospheric lifetime of minutes to hours) and is highly variable. In the minds of most climate scientists, N 2 O survives the challenge from tropospheric O 3 to its third place ranking among anthropogenic GHGs.
We conclude this section by noting the radiative forcing of climate due to tropospheric ozone is due mainly to enhancements over background levels in the tropical upper troposphere . Elevated levels of ozone in this region of the atmosphere are mainly due to biomass burning (Anderson et al. 2016). It is therefore likely that air quality regulations in the developed world will have little effect on the RF of climate due to tropospheric O 3 , since so much of the developed world is outside of the tropics. Reducing human set fires in the tropics is a vexing problem, given that many of the fires are set to clear land for agriculture. Nonetheless, the development of effective controls on human set fires will likely be necessary to reduce the RF of climate due to tropospheric O 3 (Keywood et al. 2013).

HFCs, PFCs, SF 6 and NF 3
Sulfur hexafluoride (SF 6 ) and the class of compounds called hydrofluorocarbons (HFCs) and perfluorocarbons (PFCs) often appear in the climate regulation lexicon because these compounds, along with CO 2 , CH 4 , and N 2 O, were all considered by the original Kyoto Protocol. Over the course of the Anthropocene, it is estimated that the RF of climate due to SF 6 , HFCs, and PFCs has been about 0.03 W m −2 (Other F-gases, Fig. 1.4), which is about 1 % of the total RF of climate due to all anthropogenic GHGs. Nonetheless, there is concern the RF of climate of these compounds could rise in the future (IPCC/TEAP 2005;Velders et al. 2009;Zhang et al. 2011). The Doha amendment, adopted in December 2012, added nitrogen trifluoride (NF 3 ) to the list of GHGs in the Kyoto Protocol. As such, we'll provide a brief description of the lifetimes, GWP, and industrial uses of HFCs, PFCs, SF 6 , and NF 3 .
First a little demystification of the chemistry. All of the compounds considered in this section contain at least one fluorine (F) atom, which is in the halogen column of the periodic table. Also and most importantly, none of the compounds discussed here contain any chlorine or bromine atoms. Chlorine (Cl) and bromine (Br), two other halogens, are harmful to Earth's ozone layer and any industrial compound containing either Cl or Br that has a long enough lifetime to reach the stratosphere falls under the auspices of the Montreal Protocol. Natural production of HFCs, PFCs, SF 6 and NF 3 does not occur. Therefore, the presence of these compounds in the atmosphere at a detectable level is attributed to human activity.
The HFCs, PFCs, SF 6 and NF 3 group of GHGs are chemically stable and radiatively active. Most of these compounds have either a single central element surrounded by either numerous fluorine atoms or some combination of fluorine and hydrogen atoms, or a central double carbon similarly surrounded. These chemicals have various physical properties that have resulted in a wide range of industrial applications. The molecular structure of these compounds makes them very long lived: most survive intact until they encounter the intense ultraviolet radiation environment of Earth's upper stratosphere, except for some HFCs that are removed by chemical reactions in Earth's troposphere. Finally, the presence of F in these molecules creates what scientists call a strong dipole moment. These dipole moments tend to occur at wavelengths where thermal radiation emitted by Earth's surface would otherwise escape to space (i.e., an atmospheric window). Chemicals that are long-lived and absorb in an atmospheric window tend to have large GWPs. Typically, the more F in a compound, the higher the GWP (Bera et al. 2009 (2013) and is intended to serve as a synopsis of this longer table, which spans four pages, covers more than 100 compounds, and contains many properties for each compound. Only ranges of GWPs and lifetimes are given for HFCs and PFCs in Table 1.2. The atmospheric lifetime for some of these molecules is remarkably long (CF 4 , a PFC, has a lifetime of 50,000 years) and many of the GWPs are huge (C 2 F 6 , another PFC, has a GWP of 11,100).
Hydrofluorocarbons (HFCs) reside at the intersection of ozone depletion and global warming. The Montreal Protocol, which was enacted to protect Earth's ozone layer, guided a transition from industrial production of CFCs to a class of gases called hydrochlorofluorcarbons (HCFCs), because HCFCs are less harmful to the O 3 layer than CFCs. 30 The Montreal Protocol requires a further transition from HCFCs to hydrofluorcarbons (HFCs) because, as noted above, HFCs pose no threat to the ozone layer. However, the GWPs of HFCs (Table 1.2) generally far exceed the GWPs of HCFCs (Table 8. A.1, IPCC (2013).
The future RF of climate due to HFCs is uncertain. Velders et al. (2009) project the RF of climate due to HFCs could be 0.4 W m −2 by mid-century, considerably larger than the RF due to HFCs considered by IPCC (2013). The primary reason for this difference is their projection of considerably larger growth in the atmospheric abundance of HFC-125 (formula CHF 2 CF 3 ; lifetime = 28 years; GWP = 3170) than in the scenarios used to guide the IPCC climate models.
A number of scientists and policy-makers have lobbied for HFCs to be removed from the UNFCCC basket of GHGs and placed under the auspices of the Montreal Protocol. The argument for this transition is twofold: (1) the production of HFCs was initiated by the Montreal Protocol; (2) this governing body has been extraordinarily effective due to close cooperation between atmospheric scientists, the chemical manufacturing industry, and policy members who staff the Parties of the 30 CFCs are a class of chemicals that contain chlorine, fluorine, and carbon atoms, whereas HCFCs are a class of chemicals that contain hydrogen, chlorine, fluorine, and carbon atoms. In some ways, bookkeeping would be easier had the former been labeled ClFCs and the latter HClFCs. Alas, the first "C" in these compounds stands for chlorine and the second stands for carbon. To make matters more confusing, HFCs are chemicals that contain only hydrogen, fluorine, and carbon atoms. Here the "C" stands for carbon. So if the C comes after the F, it stands for carbon. Montreal Protocol. Also, it is worth noting that it is inconceivable that the gross domestic product of any country could be adversely affected by regulation of HFCs. In other words, the stakes for the world's economies are low with regard to regulation of HFCs. On 15 October 2016, at the 28th Meeting of the Parties of the Montreal Protocol held in Kigali, Rwanda, an agreement was reached to regulate the future production of HFCs under the Montreal Protocol. This marks the first time the Montreal Protocol has had direct authority over a class of chemical compounds that pose no threat to the ozone layer. Perfluorocarbons (PFCs) are a class of compounds containing only carbon and fluorine that resist heat, oils, and staining. The most abundant PFCs are PFC-14 (CF 4 ), PFC-116 (C 2 F 6 ), and PFC-218 (C 3 F 8 ). Atmospheric levels of these compounds have risen steadily; contemporary levels of CF 4 , C 2 F 6 , and C 3 F 8 are a factor of 2, 4, and 10 larger, respectively, than observed during the onset of observations in the early 1970s . It has been projected that the RF of climate due to all PFCs could approach 0.04 W m −2 by end of this century (IPCC/TEAP 2005;Zhang et al. 2011). While this would represent only a small contribution to global warming, PFCs will continue to be monitored due to their extremely long atmospheric lifetimes (Table 1.2).
Sulfur hexafluoride (SF 6 ) is an excellent insulator favored in the high voltage, electric industry because this compound is non-flammable. 31 The atmospheric abundance of SF 6 has risen steadily since the early 1970s and shows no sign of abating ). It has been estimated that the RF of climate due to SF 6 could reach 0.037 W m −2 by the end of the century (Zhang et al. 2011). As for PFCs, SF 6 bears monitoring due to its atmospheric lifetime of 3200 years (Table 1.2).
The sulfur and fluorine compound sulfuryl fluoride (SO 2 F 2 ) is used as an insecticide and is also monitored, due to a GWP of 800. However, SO 2 F 2 has a lifetime of only 36 years. As a result, atmospheric abundances would decline relatively soon after any corrective action were taken, if such action were ever needed.
The sulfur, fluorine, and carbon containing compound SF 5 CF 3 received considerable attention in the media following discovery of a surprisingly large atmospheric abundance (Sturges et al. 2000). This gas was termed a "super GHG" because it has the highest radiative efficiency, 0.57 W m −2 ppb −1 , of any GHG ever studied. However, recent measurements reveal a slowdown in the emissions to the atmosphere (Sturges et al. 2012) and the present RF of climate of SF 5 CF 3 is a miniscule ~8.6 × 10 −5 W m −2 . 32 Finally, nitrogen trifluoride (NF 3 ) is the latest member of the GHG-club. In 2008, several studies appeared calling attention to the RF of climate due to this previously unappreciated compound (Prather and Hsu 2008;Tsai 2008). As noted above, NF 3 was added to the Kyoto Protocol list of GHGs as part of the Doha amendment in 2012. The lifetime and GWP of NF 3 are given in Table 1.2. The primary atmo-31 At one time SF 6 was used to cushion sports shoes, but this use ceased a decade ago and is not considered to be atmospherically important. 32 Atmospheric abundance of SF 5 CF 3 was 0.00015 ppb in 2012 (Sturges et al. 2012); RF of SF 5 CF 3 = 0.00015 ppb × 0.57 W m −2 ppb −1 = 8.6 × 10 −5 W m −2 . spheric release of NF 3 seems to be due to the manufacture of large, liquid crystal display screens (Thomas et al. 2012). The present RF of climate due to NF 3 is small, ~2.4 × 10 −4 W m −2 . 33 Perhaps this late-comer to the GHG-club will one day be known as the couch potato GHG.

Aerosols
Aerosols are small solid or liquid particles suspended in air. In the context of this book, we use aerosols to refer to particles either emitted directly into the atmosphere by a particular human activity (typically fossil fuel combustion or fires) or particles that form following chemical and physical transformations in the atmosphere of pollutants known as aerosol precursors. The only natural aerosols we shall consider are those resulting from volcanic eruptions; volcanic aerosols only affect climate if they exist in the stratosphere.
Aerosols, particularly those containing the element sulfur, reflect incoming solar radiation, which cools the surface. Sulfate aerosols tend to be produced from pollutants emitted by coal-fired power plants, ships, and diesel fueled trucks and cars, although there is a strong movement towards use of ultra-low sulfur diesel fuel in some parts of the world . Volcanic aerosols, which exert shortterm climatic cooling, are also composed of sulfate (Lacis and Mischenko 1995). Sooty aerosols, termed black carbon, are likewise produced by combustion of fossil fuels and biomass burning. Black carbon aerosols have a warming effect because these particles absorb solar radiation (Bond et al. 2013).
The association of human activity with the presence of tropospheric aerosols is well established from both ground-based (Jimenez et al. 2009;Yoon et al. 2016) and space-based observations (Streets et al. 2013;Yoon et al. 2014;He et al. 2016;McLinden et al. 2016). Yet, quantification of the RF of climate due to tropospheric aerosols continues to pose a scientific challenge due to the inability to precisely define numerical values of both the direct modulation of RF by anthropogenic aerosols (Myhre 2009;Kahn 2012;Bond et al. 2013) and the changes in RF driven by the effect of aerosols on clouds (Morgan et al. 2006;Carslaw et al. 2013). The IPCC (2013) best estimate and uncertainty of ΔRF over the course of the Anthropocene for these two terms, labeled Aerosol Direct Effect and Aerosol-Cloud Interaction, are shown in Fig. 1.4.
Tropospheric aerosols lie at the nexus of public health, air quality, and climate change. Exposure to small (Dominici et al. 2006) and/or toxic (Bell et al. 2007) aerosols has deleterious effects on human health. As a consequence, movements are underway throughout the world to reduce both the direct emission of aerosols as well as the emission of aerosol precursors. Reductions in the abundance of tropospheric aerosols and aerosol precursors, in response to air quality legislation motivated by public health concerns, have been readily observed by space-borne 33 Atmospheric abundance of NF 3 peaked at 0.0012 ppb in late 2011 (Arnold et al. 2012) and radiative efficiency is 0.2 W m −2 ppb −1 ( instrumentation throughout the world (Streets et al. 2013;Yoon et al. 2014;He et al. 2016). As such, the climate system is presently transitioning from an era where the cooling of climate due to aerosols may have had close to comparable strength as GHG induced warming to an era where the radiative warming due to GHGs will dominate aerosol cooling (Smith and Bond 2014).
The transition to a GHG dominated regime is illustrated in Fig. 1.10. This figure shows ΔRF due to CO 2 , CH 4 , and N 2 O as well as all anthropogenic GHGs from 1850 to 2100 for the Representative Concentration Pathway (RCP) 4.5 scenario ) used throughout IPCC (2013. Total ΔRF due to all anthropogenic GHGs reaches 4.5 W m −2 in year 2100, as designed. The error bars for the ΔRF terms of GHGs, placed at year 2011, are from IPCC (2013). These uncertainties represent 5 and 95 % confidence intervals. Figure 1.10 shows 71 plausible values for time series of ΔRF due to tropospheric aerosols published by Smith and Bond (2014). The colors correspond to least cooling (reds) to most cooling (blues); the black line denotes the central (median) scenario. These estimates are based on time series of the direct RF of climate due to black carbon, organic carbon, and sulfate aerosols as well as the effect of aerosols on clouds, all tied to the emissions of aerosols and aerosol precursors from the RCP 4.5 scenario. There exists considerable uncertainty with each of these terms. Most importantly, these uncertainties are handled in a self-consistent manner for each of the 71 scenarios over the time period 1850-2100. The scenarios colored in red (least cooling) assume black carbon aerosols exert considerable warming of climate, offsetting nearly all of the cooling by sulfate and organic carbon and the effect of aerosols of clouds. Conversely, the scenarios colored in blue (most cooling) assume black carbon aerosols exert little warming and that sulfate plus organic carbon, combined with the cloud response have led to about 1.4 W m −2 cooling in year 2011. For these large cooling scenarios, tropospheric aerosols offset nearly half of the ~2.8 W m −2 warming due to GHGs in year 2011.
The difference between the blue and red curves represents the uncertainty in the radiative forcing of climate due to aerosols. As we shall see in Chap. 2, this uncertainty limits our ability to forecast future global warming. All of the aerosol scenarios converge to near zero ΔRF in year 2100. Forecast values of ΔT in 2100 depend on ΔRF from GHGs (known well, provided CO 2 , CH 4 , N 2 O, and the minor GHGs are specified) combined with the true value of climate feedback (see Sect. 2.2.1.2). The climate record over 1850 to present can be fit nearly equally well under two contrasting scenarios: (i) the true value of aerosol RF happened to be little cooling (red curves, aerosols, Fig. 1.10) in which case climate feedback must be modest; (ii) the true value of aerosol RF happened to be large cooling (blue curves) in which case climate feedback must be considerable. If we assume the feedback inferred from the climate record persists over time, then the future rise in ΔT for the modest feedback scenario will be considerably smaller than the future rise in ΔT for the considerable feedback scenario. Even though the human fingerprint on tropospheric aerosol loading is extremely well established, uncertainty in the climatically critical quantity ΔRF due to aerosols leads to considerable spread in future projections of global warming.

Methods
Most of the figures are composites formed by combining data from publicly available data archives. Here we provide details on webpage addresses of these archives, citations to the scientific papers that describe the measurements, as well as details regarding how the data has been processed. Electronic copies of the figures are available on-line at http://parisbeaconofhope.org. Figure 1.1 shows estimates of the global mean surface temperature anomaly (ΔT) relative to the pre-industrial baseline and the mixing ratio of atmospheric CO 2 , plotted using a logarithmic scale. The figure is broken up into six intervals, denoted using Era. Sources of ΔT and CO 2 for each Era are described below.
Era 1, ΔT is based on two data records: Eras 2 and 3, ΔT is based on the European Project for Ice Coring in Antarctica (EPICA) Dome C record (Jouzel et al. 2007) archived at: http://www1.ncdc.noaa.gov/pub/data/paleo/icecore/antarctica/epica_domec/ edc3deuttemp2007.txt This record is based on analysis of the isotopic composition of the ice core, which is sensitive to temperature conditions at the time the ice formed. Two adjustments have been applied. First, we have subtracted 0.4250 °C from each data point to place the record on our 1850-1900 baseline. Second, since the ice core record represents temperature anomalies in Antarctica, which are larger than for other parts of the world, we have multiplied each data point by 0.463 to account for this difference. This multiplicative factor, based on analysis of the relation between Arctic and global warming over the modern time period (Chylek and Lohmann 2005), is in good agreement with the climate model simulations of the relation between warming in Antarctica and throughout the world (Masson-Delmotte et al. 2006).
Eras 4 and 5, ΔT is based on changes in Earth's surface temperature inferred from observations of isotopic composition of the shells preserved in deep seas cores (Hansen et al. 2013), archived at Columbia University: http://www.columbia.edu/~mhs119/Sensitivity+SL+CO2/ Table.txt Data in column 6, labeled Ts, are used. The authors have related these deep sea core inferences to 14 °C, which is the globally averaged surface temperature from 1961 to 1990. We have subtracted 14 °C from each data point to turn the record into an anomaly relative the 1961-1990 baseline, then added 0.3134 °C to each data point to place this record on our 1850-1900 baseline.
Era6, ΔT is based on the isotopic composition of marine carbonates corrected for the influence of oceanic acidity and adjusted also for modeled variations of ancient, atmospheric CO 2 (Royer et al. 2004) archived at: http://www.realclimate.org/docs/Temp-summary-from- Royer-et-al-2004.xls Data in column D of this file were used. The authors have estimated changes in deep sea temperature relative to present. We have converted to surface temperature anomaly by multiplying their record by 2.5, the ratio of changes in global surface temperature to deep sea temperature according to equation 4.2 of Hansen et al. (2013). We have interpreted present to mean the 1961-1990 baseline, so we have also added 0.3134 °C to the data so that this record is also reflective of our 1850-1900 baseline. Eras 2 and 3, CO 2 is based on a merged ice core data set that combines measurements from seven ice cores archived at: http://www1.ncdc.noaa.gov/pub/data/paleo/icecore/antarctica/antarcti-ca2015co2.xls This record is also based on laboratory measurement of CO 2 in air extracted from the ice (e.g., Petit et al. 1999). Column 2 of the CO2_Composite tab of the Excel file has been used; this composite is based on ten publications, all cited in the file.
Eras 4, 5, and 6, CO 2 is based on proxy estimates from five methods, originating from more than a hundred individual publications, summarized by Royer et al. (2012) and Peppe and Royer (2015). We have used a data file containing these observations sent to us by Dana Royer, senior author of these papers. Data for each proxy was first averaged, for all points falling within temporal bins of width 1 million years for Era 4, 5 million years for Era 5, and 50 million years for Era 5. Then, for each time bin, all available proxy means were averaged, resulting in the CO 2 time series connected by the blue lines. The error bars represent the minimum and maximum of the various proxy means available for specific time intervals. If CO 2 from a proxy was not available for a particular bin, a linear interpolation across adjacent time bins was applied, if possible. Otherwise, CO 2 from that missing proxy was treated as not available. The time ranges spanned by the five proxies are: paleosols (1-400 Mybp); alkenones (1-40 Mybp); stomata (1-400 Mypb); boron (1-15 and 35-55 Mybp); and liverworts . Finally, the paleosol record as corrected by Breecker et al. (2009) was used. Figure 1.2 shows values for the global mean surface temperature anomaly (ΔT) relative to the 1850-1900 baseline from two sources. For years prior to 1855, the proxy temperature time series of Jones and Mann (2004) was used. For 1855 to present, instrument data from HadCRUT.4.4.0.0 (Jones et al. 2012) was used. Both datasets were downloaded from the websites described for Fig. 1.1 Aerosols terms includes the direct radiative effect of aerosols, the perturbation to the reflectivity of clouds induced by aerosols, and the darkening of snow caused by the deposition of black carbon. The following types of aerosols were considered: sulfate, organic carbon and black carbon from both fossil fuel combustion and biomass burning, nitrate, and mineral dust. The total anthropogenic term combines the contributions to RF of climate from all GHGs released by human activity, plus RF of climate due to aerosols, depletion of stratospheric O 3 , the increase of tropospheric O 3 , and rising surface reflectivity due to land use change. Figure 1.3 also shows the global mean surface temperature anomaly (ΔT) relative to pre-industrial (1850-1900) baseline. Data sources for ΔT are the same as for Fig. 1.2.  (2013), "the calculation of ERF requires longer simulations with more complex models than the calculation of RF, but the inclusion of the additional rapid adjustments makes ERF a better indicator of the eventual global mean temperature response, especially for aerosols". We have used a mixture of ΔRF and ΔERF values for Fig. 1.4 because this is all that is available from Chap. 8 of IPCC (2013). Table 1.3 provides numerical estimates of the value, uncertainty, and origin of the data used in Fig. 1.4. All uncertainties represent 5-95 % confidence intervals and are given as a range, rather than a plus and minus value, since some are asymmetric about the mean. Figure 1.5 shows a profile of the change in temperature over the time period 1959-2012, based on radiosonde observations collected in the latitude range 30°S to 30°N (Sherwood and Nishant 2015). Data reflect the Iterative Universal Kriging This term considers the RF of climate due to HFCs, PFCs, SF 6 , and a few other long-lived fluorinated species. The IPCC (2013) definition combines the RF of climate due to HFC-23, HFC-32, HFC-125, HFC-134a, HFC-143a, HFC-152a, CF 4 , C 2 F 6 , SF 6 , SO 2 F 6 , and NF 3 c This term represents the RF of climate due to the increase in stratospheric H 2 O driven by rising levels of tropospheric CH 4 . It does not include radiative effects of changes in stratospheric H 2 O that occur in response to global warming (Solomon et al. 2010) (IUKv2) processing. Results are displayed as a function of altitude, rather than pressure, using a standard climatology for altitude versus pressure of the tropical atmosphere. The tropopause has been placed at the altitude corresponding to a pressure of 100 hPa. The patterns of tropospheric warming, stratospheric cooling, and drop in the tropospheric lapse rate (i.e., more warming aloft than at the surface) illustrated in Fig. 1.5 are seen throughout the global atmosphere, in addition to the tropics (Sherwood and Nishant 2015). Figure 1.6 shows CO 2 from Mauna Loa Observatory (Keeling et al. 1976) and global annual average CO 2 (Ballantyne et al. 2012) provided by NOAA ESRL at: ftp://ftp.cmdl.noaa.gov/products/trends/co2/co2_mm_mlo.txt ftp://ftp.cmdl.noaa.gov/products/trends/co2/co2_annmean_gl.txt The global CO 2 record given at the above URL starts in 1980. We have extended this record back to 1959 using annual, global average CO 2 growth rates given at: http://www.esrl.noaa.gov/gmd/ccgg/trends/global.html#global_growth The data used to construct the CO 2 emissions from the combustion of fossil fuel (Boden et al. 2013) plus land use change (Houghton et al. 2012) (green bars, Fig. 1.6a) originate from file: http://cdiac.ornl.gov/ftp/Global_Carbon_Project/Global_Carbon_Budget_ 2015_v1.1.xlsx hosted by the Carbon Dioxide Information Analysis Center (CDIAC) at the US Department of Energy's (DOE) Oak Ridge National Laboratory (ORNL). This same file is also provided by the Global Carbon Budget at: http://www.globalcarbonproject.org/carbonbudget/15/files/Global_ Carbon_Budget_2015v1.1.xlsx Contents of this file, which contains much more information than used here, are described by Le Quéré et al. (2015). The blue bars are found by multiplying the difference in annual average CO 2 mixing ratio, units of ppm, by 7.768, to arrive at the mass of CO 2 in Gt (see Le Quéré et al. (2015)). Finally, the Tropical Pacific ENSO index represents the anomaly of sea surface temperature in the region bounded by 20°S to 20°N latitude and 160°E to 80°W longitude, relative to a long-term climatology. Monthly values of this index have been computed as described by Zhang et al. (1997), using HadSST3.1.1.0 sea surface temperature data (Kennedy et al. 2011a, b) (Keeling et al. 1976), the O 2 /N 2 ratio (Keeling et al. 1996), and δ 13 C of CO 2 (Keeling et al. 2005) from Mauna Loa Observatory (MLO) as well as CO 2 from the South Pole (SPO) (Tans et al. 1990). For CO 2 , the solid black line shows monthly mean data from NOAA ESRL, based on the same file given in Methods for Fig. 1.6. Daily measurements of CO 2 at MLO (dots) are based on data provided by the Scripps Institution of Oceanography (SIO) of the University of California, San Diego at: http://scrippsco2.ucsd.edu/assets/data/atmospheric/stations/flask_co2/daily/ daily_flask_co2_mlo.csv Monthly mean CO 2 at SPO (red line) is based on data provided by NOAA ESRL at: http://scrippsco2.ucsd.edu/assets/data/atmospheric/stations/flask_co2/monthly/ monthly_flask_co2_spo.csv Monthly (line) and daily (dots) observations of the O 2 /N 2 ratio are based on data archived by SIO at: http://scrippso2.ucsd.edu/sites/default/files/data/o2_data/o2_monthly/mloo.txt http://scrippso2.ucsd.edu/sites/default/files/data/o2_data/o2_daily/mlooav.csv Finally, monthly (line) and daily (dots) 13 δCO 2 is also based on data from SIO, at: http://scrippsco2.ucsd.edu/assets/data/atmospheric/stations/flask_isotopic/ monthly/monthly_flask_c13_mlo.csv http://scrippsco2.ucsd.edu/assets/data/atmospheric/stations/flask_isotopic/daily/ daily_flask_c13_mlo.csv Figure 1.8 shows the difference between annual average CO 2 at MLO and CO 2 at SPO versus total anthropogenic emissions of CO 2 . Here, we formed annual average CO 2 for each station from monthly mean values, based on in situ and flask sampling archived by SIO at: http://scrippsco2.ucsd.edu/assets/data/atmospheric/stations/in_situ_co2/ monthly/monthly_in_situ_co2_mlo.csv http://scrippsco2.ucsd.edu/assets/data/atmospheric/stations/flask_co2/ monthly/monthly_flask_co2_spo.csv The total anthropogenic emissions of CO 2 are annual tabulations, reflecting the sum of combustion of fossil fuel (Boden et al. 2013) plus land use change (Houghton et al. 2012), and originate from the Global Carbon Project archive described in Methods for Fig. 1.6.  Table 1 of Kirschke et al. (2013). The apportionment of the human sources is based on Fig. 1 of Conrad (2009). Numerical values of the total sink (550 Tg year −1 ) and the sink due to soils (32 Tg year −1 ) are also based on the top-down estimates from table 1 of Kirschke et al. (2013), whereas the sinks due to chemical loss via reactions with tropospheric OH (452.8 Tg year −1 ), reactions with tropospheric Cl (21.4 Tg year −1 ), and stratospheric chemistry (43.7 Tg year −1 ) are based by scaling the bottom-up estimates of these quantities given in Table 1 of Kirschke et al. (2013) for the decade 2000-2009 by the ratio 518/604 = 0.8576, where 518 Tg year −1 is the total chemical sink found using the trop-down approach and 604 Tg year −1 is the total chemical sink found using the bottom-up approach. The numbers have been combined in this manner to provide a self-consistent estimate of the CH 4 source and sink terms by combining best available information from several studies and approaches. Figure 1.10 shows ΔRF due to CO 2 , CH 4 , N 2 O, and all anthropogenic GHGs from the RCP 4.5 scenario (Meinshausen et al. 2011). The numerical values have been obtained from the same file used to find ΔRF for Fig. 1.3

. The figure also
Open Access This chapter is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, duplication, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, a link is provided to the Creative Commons license and any changes made are indicated.
The images or other third party material in this chapter are included in the work's Creative Commons license, unless indicated otherwise in the credit line; if such material is not included in the work's Creative Commons license and the respective action is not permitted by statutory regulation, users will need to obtain permission from the license holder to duplicate, adapt or reproduce the material.