Climate Change - Oxford 2023

Oxford Open Climate Change, 2023, 3(1), kgad008
https://doi.org/10.1093/oxfclm/kgad008
Perspective article
Global warming in the pipeline

James E. Hansen1�, Makiko Sato1, Leon Simons2, Larissa S. Nazarenko3,4, Isabelle Sangha1, Pushker Kharecha1,
James C. Zachos5, Karina von Schuckmann6, Norman G. Loeb7, Matthew B. Osman8, Qinjian Jin9, George Tselioudis3, Eunbi Jeong10,
Andrew Lacis3, Reto Ruedy3,11, Gary Russell3, Junji Cao12, Jing Li13
1
Climate Science, Awareness and Solutions, Columbia University Earth Institute, New York, NY, USA
2
The Club of Rome Netherlands, ‘s-Hertogenbosch, The Netherlands
Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

3
NASA Goddard Institute for Space Studies, New York, NY, USA
4
Center for Climate Systems Research, Columbia University Earth Institute, New York, NY, USA
5
Earth and Planetary Science, University of CA, Santa Cruz, CA, USA
6
Mercator Ocean International, Ramonville St., -Agne, France
7
NASA Langley Research Center, Hampton, VA, USA
8
Department of Geosciences, University of AZ, Tucson, AZ, USA
9
Department of Geography and Atmospheric Science, University of KS, Lawrence, KS, USA
10
CSAS KOREA, Goyang, Gyeonggi-do, South Korea
11
Business Integra, Inc, New York, NY, USA
12
Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
13
Department of Atmospheric and Oceanic Sciences, School of Physics, Peking University, Beijing, China
�Correspondence address. Director of Climate Science, Awareness and Solutions, Earth Institute, Columbia University, 475 Riverside Drive, Ste. 401-O, New York,
NY 10115, USA. E-mail: [email protected]
Abstract
Improved knowledge of glacial-to-interglacial global temperature change yields Charney (fast-feedback) equilibrium climate sensi
tivity 1.2 ± 0.3� C (2r) per W/m2, which is 4.8� C ± 1.2� C for doubled CO2. Consistent analysis of temperature over the full Cenozoic
era—including ‘slow’ feedbacks by ice sheets and trace gases—supports this sensitivity and implies that CO2 was 300–350 ppm in the
Pliocene and about 450 ppm at transition to a nearly ice-free planet, exposing unrealistic lethargy of ice sheet models. Equilibrium
global warming for today’s GHG amount is 10� C, which is reduced to 8� C by today’s human-made aerosols. Equilibrium warming is
not ‘committed’ warming; rapid phaseout of GHG emissions would prevent most equilibrium warming from occurring. However, de
cline of aerosol emissions since 2010 should increase the 1970–2010 global warming rate of 0.18� C per decade to a post-2010 rate of at
least 0.27� C per decade. Thus, under the present geopolitical approach to GHG emissions, global warming will exceed 1.5� C in the
2020s and 2� C before 2050. Impacts on people and nature will accelerate as global warming increases hydrologic (weather) extremes.
The enormity of consequences demands a return to Holocene-level global temperature. Required actions include: (1) a global increas
ing price on GHG emissions accompanied by development of abundant, affordable, dispatchable clean energy, (2) East-West coopera
tion in a way that accommodates developing world needs, and (3) intervention with Earth’s radiation imbalance to phase down
today’s massive human-made ‘geo-transformation’ of Earth’s climate. Current political crises present an opportunity for reset, espe
cially if young people can grasp their situation.
Keywords: Aerosols; Climate Sensitivity; Paleoclimate; Global Warming; Energy Policy; Cenozoic
Background information and structure intermediate waters of the ocean to absorb heat could delay the
of paper estimated warming by several decades.’
After U.S. President Jimmy Carter signed the 1980 Energy
It has been known since the 1800s that infrared-absorbing
Security Act, which included a focus on unconventional fossil
(greenhouse) gases (GHGs) warm Earth’s surface and that the
fuels such as coal gasification and rock fracturing (‘fracking’) to
abundance of GHGs changes naturally as well as from human
extract shale oil and tight gas, the U.S. Congress asked the
actions [1, 2].1 Roger Revelle wrote in 1965 that we are conducting
National Academy of Sciences again to assess potential climate
a ‘vast geophysical experiment’ by burning fossil fuels that accu effects. Their massive Changing Climate report had a measured
mulated in Earth’s crust over hundreds of millions of years [3] tone on energy policy—amounting to a call for research [5]. Was
Carbon dioxide (CO2) in the air is now increasing and already has not enough known to caution lawmakers against taxpayer sub
reached levels that have not existed for millions of years, with sidy of the most carbon-intensive fossil fuels? Perhaps the equa
consequences that have yet to be determined. Jule Charney led a nimity was due in part to a major error: the report assumed that
study in 1979 by the United States National Academy of Sciences the delay of global warming caused by the ocean’s thermal iner
that concluded that doubling of atmospheric CO2 was likely to tia is 15 years, independent of climate sensitivity. With that as
cause global warming of 3 ± 1.5� C [4]. Charney added: ‘However, sumption, they concluded that climate sensitivity for 2 � CO2 is
we believe it is quite possible that the capacity of the near or below the low end of Charney’s 1.5–4.5� C range. If climate
Received: December 08, 2022. Revised: August 22, 2023. Accepted: September 15, 2023
# The Author(s) 2023. Published by Oxford University Press.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which
permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
2 | Oxford Open Climate Change, 2023, Vol. 3, No. 1
sensitivity was low and the lag between emissions and climate climate sensitivity (ECS) based on improved paleoclimate data
response was only 15 years, climate change would not be nearly and introduces Earth system sensitivity (ESS), which includes the
the threat that it is. feedbacks that Charney held fixed. Climate response time section
Simultaneous with preparation of Changing Climate, climate explores the fast-feedback response time of Earth’s temperature
sensitivity was addressed at the 1982 Ewing Symposium at the and energy imbalance to an imposed forcing, concluding that
Lamont Doherty Geophysical Observatory of Columbia cloud feedbacks buffer heat uptake by the ocean, thus increasing
University on 25–27 October, with papers published in January the delay in surface warming and making Earth’s energy imbal
1984 as a monograph of the American Geophysical Union [6]. ance an underestimate of the forcing reduction required to stabi
Paleoclimate data and global climate modeling together led to an lize climate. Cenozoic era section analyzes temperature change
inference that climate sensitivity is in the range 2.5–5� C for 2 � of the past 66 million years and infers the Cenozoic history of
CO2 and that climate response time to a forcing is of the order of CO2, thus providing insights about climate change. Aerosols sec
a century, not 15 years [7]. Thus, the concept that a large amount tion addresses the absence of aerosol forcing data via inferences
of additional human-made warming is already ‘in the pipeline’ from paleo data and modern global temperature change, and we

was introduced. E.E. David, Jr, President of Exxon Research and point out potential information in ‘the great inadvertent aerosol
Engineering, in his keynote talk at the symposium insightfully experiment’ provided by recent restrictions on fuels in interna
noted [8]: ‘The critical problem is that the environmental impacts tional shipping. Summary section discusses policy implications
of the CO2 buildup may be so long delayed. A look at the theory of high climate sensitivity and the delayed response of the cli
of feedback systems shows that where there is such a long delay, mate system. Warming in the pipeline need not appear. We can
the system breaks down, unless there is anticipation built into take actions that slow and reverse global warming; indeed, we
the loop.’ suggest that such actions are needed to avoid disastrous conse
Thus, the danger caused by climate’s delayed response and quences for humanity and nature. Reduction of greenhouse gas
the need for anticipatory action to alter the course of fossil fuel emissions as rapidly as practical has highest priority, but that
development was apparent to scientists and the fossil fuel indus policy alone is now inadequate and must be complemented by
try 40 years ago.2 Yet industry chose to long deny the need to additional actions to affect Earth’s energy balance. The world is
change energy course [9], and now, while governments and fi still early in this ‘vast geophysical experiment’—as far as conse
nancial interests connive, most industry adopts a ‘greenwash’ quences are concerned—but time has run short for the
approach that threatens to lock in perilous consequences for hu ‘anticipation’ that E.E. David recommended.
manity. Scientists will share responsibility if we allow govern
ments to rely on goals for future global GHG levels, as if targets
had meaning in the absence of policies required to achieve them. Climate sensitivity (ECS and ESS)
The Intergovernmental Panel on Climate Change (IPCC) was This section gives a brief overview of the history of ECS estimates
established in 1988 to provide scientific assessments on the state since the Charney report and uses glacial-to-interglacial climate
of knowledge about climate change [10] and almost all nations change to infer an improved estimate of ECS. We discuss how
agreed to the 1992 United Nations Framework Convention on ECS and the more general Earth system sensitivity (ESS) depend
Climate Change [11] with the objective to avert ‘dangerous an on the climate state.
thropogenic interference with the climate system’. The current Charney defined ECS as the eventual global temperature
IPCC Working Group 1 report [12] provides a best estimate of 3� C change caused by doubled CO2 if ice sheets, vegetation and long-
for equilibrium global climate sensitivity to 2 � CO2 and lived GHGs are fixed (except the specified CO2 doubling). Other
describes shutdown of the overturning ocean circulations and quantities affecting Earth’s energy balance—clouds, aerosols,
large sea level rise on the century time scale as ‘high impact, low water vapor, snow cover and sea ice—change rapidly in response
probability’ even under extreme GHG growth scenarios. This con to climate change. Thus, Charney’s ECS is also called the ‘fast-
trasts with ‘high impact, high probability’ assessments reached feedback’ climate sensitivity. Feedbacks interact in many ways,
in a paper [13]—hereafter abbreviated Ice Melt—that several of so their changes are calculated in global climate models (GCMs)
us published in 2016. Recently, our paper’s first author (JEH) de that simulate such interactions. Charney implicitly assumed that
scribed a long-time effort to understand the effect of ocean mix change of the ice sheets on Greenland and Antarctica—which we
ing and aerosols on observed and projected climate change, categorize as a ‘slow feedback’—was not important on time
which led to a conclusion that most climate models are unrealis scales of most public interest.
tically insensitive to freshwater injected by melting ice and that ECS defined by Charney is a gedanken concept that helps us
ice sheet models are unrealistically lethargic in the face of rapid, study the effect of human-made and natural climate forcings. If
large climate change [14]. knowledge of ECS were based only on models, it would be diffi
Eelco Rohling, editor of Oxford Open Climate Change, invited cult to narrow the range of estimated climate sensitivity—or
a perspective article on these issues. Our principal motivation in have confidence in any range—because we do not know how well
this paper is concern that IPCC has underestimated climate sen feedbacks are modeled or if the models include all significant
sitivity and understated the threat of large sea level rise and real-world feedbacks. Cloud and aerosol interactions are com
shutdown of ocean overturning circulations, but these issues, be plex, e.g. and even small cloud changes can have a large effect.
cause of their complexity, must be addressed in two steps. Our Thus, data on Earth’s paleoclimate history are essential, allowing
present paper addresses climate sensitivity and warming in the us to compare different climate states, knowing that all feed
pipeline, concluding that these exceed IPCC’s best estimates. backs operated.
Response of ocean circulation and ice sheet dynamics to global
warming—already outlined in the Ice Melt paper—will be Climate sensitivity estimated at the 1982
addressed further in a later paper. Ewing Symposium
The structure of our present paper is as follows. Climate sensi Climate sensitivity was addressed in our paper [7] for the Ewing
tivity section makes a fresh evaluation of Charney’s equilibrium Symposium monograph using the feedback framework implied
J. E. Hansen et al. | 3
by E.E. David and employed by electrical engineers [15]. The cli cloud amount, despite increases in cloud albedo [22]. These cloud
mate forcing caused by 2 � CO2—the imposed perturbation of changes are found in all observed cloud regimes and locations,
Earth’s energy balance—is �4 W/m2. If there were no climate implying robust thermodynamic control [23].
feedbacks and Earth radiated energy to space as a perfect black CLIMAP SSTs were a more likely cause of the planetary energy
surface, Earth’s temperature would need to increase �1.2� C to in imbalance. Co-author D. Peteet used pollen data to infer LGM
crease radiation to space 4 W/m2 and restore energy balance. tropical and subtropical cooling 2–3� C greater than in a GCM
However, feedbacks occur in the real world and in GCMs. In our forced by CLIMAP SSTs. D. Rind and Peteet found that montane
GCM the equilibrium response to 2 � CO2 was 4� C warming of LGM snowlines in the tropics descended 1 km in the LGM, incon
Earth’s surface. Thus, the fraction of equilibrium warming due sistent with climate constrained by CLIMAP SSTs. CLIMAP as
directly to the CO2 change was 0.3 (1.2� C/4� C) and the feedback sumed that tiny shelled marine species migrate to stay in a
‘gain’, g, was 0.7 (2.8� C/4� C). Algebraically, ECS and feedback gain temperature zone they inhabit today. But what if, instead, these
are related by species partly adapt over millennia to changing temperature?
Based on the work of Rind and Peteet, later published [24], we

ECS ¼ 1:2� C=ð1 gÞ (1) suspected but could not prove that CLIMAP SSTs were too warm.
Based on GCM simulations for 2 � CO2, on our feedback analy
We evaluated contributions of individual feedback processes to g sis for the LGM, and on observed global warming in the past cen
by inserting changes of water vapor, clouds, and surface albedo tury, we concluded that ECS was in the range 2.5–5� C for 2 � CO2.
(reflectivity, literally whiteness, due to sea ice and snow changes) If CLIMAP SSTs were accurate, ECS was near the low end of that
from the 2 � CO2 GCM simulation one-by-one into a one- range. In contrast, our analysis implied that ECS for 2 � CO2 was
dimensional radiative-convective model [16], finding gwv ¼ 0.4, in the upper half of the 2.5–5� C range, but our analysis depended
gcl ¼ 0.2, gsa ¼ 0.1, where gwv, gcl, and gsa are the water vapor, in part on our GCM, which had sensitivity 4� C for 2 � CO2. To re
cloud and surface albedo gains. The 0.2 cloud gain was about solve the matter, a paleo thermometer independent of biologic
equally from a small increase in cloud top height and a small de adaptation was needed. Several decades later, such a paleo ther
crease in cloud cover. These feedbacks all seemed reasonable, mometer and advanced analysis techniques exist. We will use re
but how could we verify their magnitudes or the net ECS due to cent studies to infer our present best estimates for ECS and ESS.
all feedbacks? First, however, we will comment on other estimates of climate
We recognized the potential of emerging paleoclimate data. sensitivity and clarify the definition of climate forcings that
Early data from polar ice cores revealed that atmospheric CO2 we employ.
was much less during glacial periods and the CLIMAP project [17]
used proxy data to reconstruct global surface conditions during IPCC and independent climate
the Last Glacial Maximum (LGM), which peaked about 20 000 sensitivity estimates
years ago. A powerful constraint was the fact that Earth had to Reviews of climate sensitivity are available, e.g. Rohling et al. [25],
be in energy balance averaged over the several millennia of the which focuses on the physics of the climate system, and
LGM. However, when we employed CLIMAP boundary conditions Sherwood et al. [26], which adds emphasis on probabilistic combi
including sea surface temperatures (SSTs), Earth was out of en nation of multiple uncertainties. Progress in narrowing the un
ergy balance, radiating 2.1 W/m2 to space, i.e. Earth was trying to certainty in climate sensitivity was slow in the first five IPCC
cool off with an enormous energy imbalance, equivalent to half assessment reports. The fifth assessment report [26] (AR5) in
of 2 � CO2 forcing. 2014 concluded only—with 66% probability—that ECS was in the
Something was wrong with either assumed LGM conditions or range 1.5–4.5� C, the same as Charney’s report 35 years earlier.
our climate model. We tried CLIMAP’s maximal land ice—this The broad spectrum of information on climate change—espe
only reduced the energy imbalance from 2.1 to 1.6 W/m2. cially constraints imposed by paleoclimate data—at last affected
Moreover, we had taken LGM CO2 as 200 ppm and did not know AR6 [12], which concluded with 66% probability that ECS is 2.5–
that CH4 and N2O were less in the LGM than in the present inter 4� C, with 3� C as their best estimate (Supplementary Fig. TS.6).
glacial period; accurate GHGs and CLIMAP SSTs produce a plane Sherwood et al. [21] combine three lines of evidence: climate
tary energy imbalance close to 3 W/m2. Most feedbacks in our feedback studies, historical climate change, and paleoclimate
model were set by CLIMAP. Sea ice is set by CLIMAP. Water vapor data, inferring S ¼ 2.6–3.9� C with 66% probability for 2 � CO2,
depends on surface temperature, which is set by CLIMAP SSTs. where S is an ‘effective sensitivity’ relevant to a 150-year time
Cloud feedback is uncertain, but ECS smaller than 2.4� C for 2 � scale. They find ECS only slightly larger: 2.6–4.1� C with 66% prob
CO2 would require a negative cloud gain. gcl � 0.2 from our GCM ability. Climate feedback studies, inherently, cannot yield a sharp
increases ECS from 2.4� C to 4� C (Equation 1) and accounts for al definition of ECS, as we showed in the cloud feedback discussion
most the entire difference of sensitivities of our model (4� C for 2 above. Earth’s climate system includes amplifying feedbacks that
� CO2) and the Manabe and Stouffer model [18] (2� C for 2 � CO2) push the gain, g, closer to unity than zero, thus making ECS sen
that had fixed cloud cover and cloud height. Manabe suggested sitive to uncertainty in any feedback; the resulting sensitivity of
[19] that our higher ECS was due to a too-large sea ice and snow ECS to g prohibits precise evaluation from feedback analysis.
feedback, but we noted [7] that sea ice in our control run was less Similarly, historical climate change cannot define ECS well be
than observed, so we likely understated sea ice feedback. cause the aerosol climate forcing is unmeasured. Also, forced
Amplifying feedback due to high clouds increasing in height with and unforced ocean dynamics give rise to a pattern effect: [27]
warming is expected and is found in observations, large-eddy the geographic pattern of transient and equilibrium temperature
simulations and GCMs [20] Sherwood et al. [21] conclude that changes differ, which affects ECS inferred from transient climate
negative low-cloud feedback is ‘neither credibly suggested by any change. These difficulties help explain how Sherwood et al. [21]
model, nor by physical principles, nor by observations.’ Despite a could estimate ECS as only 6% larger than S, an implausible re
wide spread among models, GCMs today show an amplifying sult in view of the ocean’s great thermal inertia. An intercompar
cloud feedback due to increases in cloud height and decreases in ison of GCMs run for millennial time scales, LongRunMIP [28],
includes 14 simulations of 9 GCMs with runs of 5000 years Fs, not Fo. It would be useful if both Fo and dTo were reported for
(or close enough for extrapolation to 5000 years). Their global all climate models.
warmings at 5000 years range from 30% to 80% larger than their A further refinement of climate forcing is suggested in Efficacy:
150-year responses. effective forcing (Fe) defined by a long GCM run with calculated
Our approach is to compare glacial and interglacial equilib ocean temperature. The resulting global surface temperature
rium climate states. The change of atmospheric and surface forc change, relative to that for equal CO2 forcing, defines the forc
ings can be defined accurately, thus leading to a sharp ing’s efficacy. Effective forcings, Fe, were found to be within a few
evaluation of ECS for cases in which equilibrium response is as percent of Fs for most forcing agents, i.e. the results confirm that
sured. With this knowledge in hand, additional information can Fs is a robust forcing. This support is for Fs, not for Fo ¼ ERF,
be extracted from historical and paleo climate changes. which is systematically smaller than Fs. The Goddard Institute
for Space Studies (GISS) GCM [32, 33] used for CMIP6 [34] studies,
Climate forcing definitions which we label the GISS (2020) model,3 has higher resolution
Attention to climate forcing definitions is essential for quantita (2� � 2.5� and 40 atmospheric layers) and other changes that

tive analysis of climate change. However, readers uninterested in yield a moister upper troposphere and lower stratosphere, rela
radiative forcings may skip this section with little penalty. We de tive to the GISS model used in Efficacy. GHG forcings reported for
scribe our climate forcing definition and compare our forcings the GISS (2020) model [32, 33] are smaller than in prior GISS mod
with those of IPCC. Our total GHG forcing matches that of IPCC els, a change attributed [33] to blanketing by high level water va
within a few percent, but this close fit hides larger differences in por. However, part of the change is from comparison of Fo in GISS
individual forcings that deserve attention. (2020) to FS in earlier models. The 2 � CO2 fixed SST simulation
Equilibrium global surface temperature change is related to with the GISS (2020) model yields Fo ¼ 3.59 W/m2, dTo ¼ 0.27� C
ECS by and k ¼ 0.9� C per W/m2. Thus FS ¼ 3.59 þ 0.30 ¼ 3.89 W/m2,
which is only 5.4% smaller than the FS ¼ 4.11 W/m2 for the GISS
DTs � F � ECS ¼ F � k; (2) model used in Efficacy.
Our GHG effective forcing, Fe, was obtained in two steps.
where k is a widely used abbreviation of ECS, DTS is the global Adjusted forcings, Fa, were calculated for each gas for a large
mean equilibrium surface temperature change in response to cli range of gas amount with a global-mean radiative-convective
mate forcing F, which is measured in W/m2 averaged over the en model that incorporated the GISS GCM radiation code, which
tire planetary surface. There are alternative ways to define F, as uses the correlated k-distribution method [35] and high spectral
discussed in Chapter 8 [29] of AR5 and in a paper [30] hereafter resolution laboratory data [36]. The Fa are converted to effective
called Efficacy. Objectives are to find a definition of F such that forcings (Fe) via efficacy factors (Ea; Table 1 of Efficacy) based on
different forcing mechanisms of the same magnitude yield a sim GCM simulations that include the 3-D distribution of each gas.
ilar global temperature change, but also a definition that can be The total GHG forcing is
computed easily and reliably. The first four IPCC reports used ad
justed forcing, Fa, which is Earth’s energy imbalance after strato Fe ¼ Fa ðCO2 Þ þ 1:45Fa ðCH4 Þ þ 1:04Fa ðN2 OÞ
spheric temperature adjusts to presence of the forcing agent. Fa þ 1:32Fa ðMPTGsþOTGsÞ þ 0:45Fa ðO3 Þ (4)
usually yields a consistent response among different forcing
agents, but there are exceptions such as black carbon aerosols; Fa The CH4 coefficient (1.45) includes the effect of CH4 on O3 and
exaggerates their impact. Also, Fa is awkward to compute and stratospheric H2O, as well as the efficacy (1.10) of CH4 per se. We
depends on definition of the tropopause, which varies among assume that CH4 is responsible for 45% of the O3 change [37].
models. Fs, the fixed SST forcing (including fixed sea ice), is more Forcing caused by the remaining 55% of the O3 change is based
robust than Fa as a predictor of climate response [30, 31], but a on IPCC AR6 O3 forcing (Fa ¼ 0.47 W/m2 in 2019); we multiply this
GCM is required to compute Fs. In Efficacy, Fs is defined as AR6 O3 forcing by 0.55 � 0.82 ¼ 0.45, where 0.82 is the efficacy of
O3 forcing from Table 1 of Efficacy. Thus, the non-CH4 portion of
dTo the O3 forcing is 0.21 W/m2 in 2019. MPTGs and OTGs are
Fs ¼ Fo þ ; (3)
k Montreal Protocol Trace Gases and Other Trace Gases [38]. A list
of these gases and a table of annual forcings since 1992 are avail
where Fo is Earth’s energy imbalance after atmosphere and land
able as well as the earlier data [39].
surface adjust to the presence of the forcing agent with SST fixed.
The climate forcing from our formulae is slightly larger than
Fo is not a full measure of the strength of a forcing, because a
IPCC AR6 forcings (Fig. 1). In 2019, the final year of AR6 data, our
portion (dTo) of the equilibrium warming is already present as Fo
GHG forcing is 4.00 W/m2; the AR6 forcing is 3.84 W/m2. Our forc
is computed. A GCM run of about 100 years is needed to accu
ing should be larger, because IPCC forcings are Fo for all gases ex
rately define Fo because of unforced atmospheric variability.
cept O3, for which they provide Fa (AR6 section 7.3.2.5). Table 1 in
That GCM run also defines dTo, the global mean surface air tem
perature change caused by the forcing with SST fixed. k is the
Table 1. Greenhouse gas radiative forcings
model’s ECS in � C per W/m2. dTo/k is the portion of the total forc
ing (Fs) that is ‘used up’ in causing the dTo warming; radiative Gas Radiative Forcing
flux to space increases by dTo/k due to warming of the land sur
CO2 F ¼ f(c) – f(co), where f(c) ¼ 4.996 ln (c þ 0.0005c2)
face and global air. The term dTo/k is usually, but not always, less
CH4 F ¼ 0.0406(�m − �mo) − [g(m, no) – g(mo, no)]
than 10% of Fo. Thus, it is better not to neglect dTo/k. IPCC AR5 N 2O F ¼ 0.136(�n – �no) – [g(mo, n) – g(mo, no)]
and AR6 define effective radiative forcing as ERF ¼ Fo. Omission where g(m, n) ¼ 0.5 ln [1 þ 2 � 10–5(mn)0.75]
of dTo/k was intentional [29] and is not an issue if the practice is CFC-11 F ¼ 0.264(x − xo)
CFC-12 F ¼ 0.323(y − yo)
followed consistently. However, when the forcing is used to cal
culate global surface temperature response, the forcing to use is c, CO2 (ppm); m, CH4 (ppb); n, N2O (ppb); x/y, CFC-11/12 (ppb).
Efficacy allows accurate comparison: dTo for 2 � CO2 for the GISS not mean that CO2 is the primal cause of the climate oscillations.
model used in Efficacy is 0.22� C, k is 0.67� C per W/m2, so dTo/k ¼ Hays et al. [42] showed that small changes of Earth’s orbit and the
0.33 W/m2. Thus, the conversion factor from Fo to Fe (or Fs) is tilt of Earth’s spin axis are pacemakers of the ice ages. Orbital
4.11/(4.11−0.33). The non-O3 portion of AR6 2019 forcing changes alter the seasonal and geographical distribution of inso
(3.84−0.47 ¼ 3.37) W/m2 increases to 3.664 W/m2. The O3 portion lation, which affects ice sheet size and GHG amount. Long-term
of the AR6 2019 forcing (0.47 W/m2) decreases to 0.385 W/m2 be climate is sensitive because ice sheets and GHGs act as amplify
cause the efficacy of Fa(O3) is 0.82. The AR6 GHG forcing in 2019 ing feedbacks: [43] as Earth warms, ice sheets shrink, expose a
is thus �4.05 W/m2, expressed as Fe � Fs, which is �1% larger darker surface, and absorb more sunlight; also, as Earth warms,
than follows from our formulae. This precise agreement is not in the ocean and continents release GHGs to the air. These amplify
dicative of the true uncertainty in the GHG forcing, which IPCC ing feedbacks work in the opposite sense as Earth cools. Orbital
AR6 estimates as 10%, thus about 0.4 W/m2. We concur with forcings oscillate slowly over tens and hundreds of thousands of
their error estimate and employ it in our ECS uncertainty analy years [44]. The picture of how Earth orbital changes drive millen
sis (Equilibrium climate sensitivity section). nial climate change was painted in the 1920s by Milutin

We conclude that the GHG increase since 1750 already produ Milankovitch, who built on 19th century hypotheses of James
ces a climate forcing equivalent to that of 2 � CO2 (our formulae Croll and Joseph Adhe �mar. Paleoclimate changes of ice sheets
yield Fe � Fs ¼ 4.08 W/m2 for 2021 and 4.13 W/m2 for 2022; IPCC and GHGs are sometimes described as slow feedbacks [45], but
AR6 has Fs ¼ 4.14 W/m2 for 2021). The human-made 2 � CO2 cli their slow change is paced by the Earth orbital forcing; their slow
mate forcing imagined by Charney, Tyndall and other green change does not mean that these feedbacks cannot operate more
house giants is no longer imaginary. Humanity is now taking its rapidly in response to a rapid climate forcing.
first steps into the period of consequences. Earth’s paleoclimate We evaluate ECS by comparing stable climate states before
history helps us assess the potential outcomes. and after a glacial-to-interglacial climate transition. GHG
amounts are known from ice cores and ice sheet sizes are known
Glacial-to-interglacial climate oscillations from geologic data. This empirical ECS applies to the range of
In this section we describe how ice core data help us assess ECS global temperature covered by ice cores, which we will conclude
for climate states from glacial conditions to interglacial periods is about –7� C to þ 1� C relative to the Holocene. The Holocene is
such as the Holocene, the interglacial period of the past 12 000 an unusual interglacial. Maximum melt rate was at 13.2 kyBP, as
years. We discuss climate sensitivity in warmer climates in expected [45] and GHG amounts began to decline after peaking
Cenozoic era section. early in the Holocene, as in most interglacials. However, several
Air bubbles in Antarctic ice cores—trapped as snow piled up ky later, CO2 and CH4 increased, raising a question of whether
and compressed into ice—preserve a record of long-lived GHGs humans were affecting GHGs. Ruddiman [46] suggests that defor
for at least 800 000 years. Isotopic composition of the ice provides estation began to affect CO2 6500 years ago and rice irrigation be
a measure of temperature in and near Antarctica [40]. Changes gan to affect CH4 5000 years ago. Those possibilities complicate
of temperature and CO2 are highly correlated (Fig. 2). This does use of LGM-Holocene warming to estimate ECS. However, sea
level, and thus the size of the ice sheets, had stabilized by 7000
years ago (Evidence of aerosol forcing in the Holocene section).
Total Greenhouse Gas Forcing Thus, the millennium centered on 7 kyBP provides a good period
4 to compare with the LGM. Comparison of the Eemian interglacial
IPCC AR6 Table AIII.3
(Fa for O3, Fo for other gases) (Fig. 2) with the prior glacial maximum (PGM) has potential for in
Equation (4) dependent assessment.
3
LGM-Holocene and PGM-Eemian evaluation
F (W/m2)
of ECS
2
In this section we evaluate ECS by comparing neighboring glacial
and interglacial periods when Earth was in energy balance within
1 less than 0.1 W/m2 averaged over a millennium. Larger imbal
ance would cause temperature or sea level change that did not
occur [48].4 Thus, we can assess ECS from knowledge of atmo
0 spheric and surface forcings that maintained these climates.
1750 1800 1850 1900 1950 2000
Recent advanced analysis techniques allow improved estimate
Figure 1. IPCC AR6 Annex III greenhouse gas forcing [12], which employs of paleo temperatures. Tierney et al. [49] exclude microbiology fos
Fa for O3 and Fo for other GHGs, compared with the effective forcing, Fe, sils whose potential to adapt makes them dubious thermometers.
from Equation (4). See discussion in text. Instead, they use a large collection of geochemical (isotope)
CO2 (ppm)
5
Temperature (ºC)
Temperature 300
CO2 Eemian
0 Holocene
250
−5
200
−10
800 700 600 500 400 300 200 100 0
Date (kyBP)
Figure 2. Antarctic Dome C temperature for past 800 ky from Jouzel et al. [40] relative to the mean of the last 10 ky and Dome C CO2 amount from Luthi
et al. [41] (kyBP is kiloyears before present).
proxies for SST in an analysis constrained by climate change pat ice sheets, (2) vegetation change, e.g. boreal forests replaced by
terns defined by GCMs. They find cooling of 6.1� C (95% confi brighter tundra, and (3) continental shelves exposed by lower sea
dence: level. Forcing by all three can be evaluated at once with a GCM.
5.7–6.5� C) for the interval 23–19 kyBP. A similarly constrained Accuracy requires realistic clouds, which shield the surface.
global analysis by Osman et al. [50] finds LGM cooling at 21–18 Clouds are the most uncertain feedback [52]. Evaluation is ideal
kyBP of 7.0 ± 1� C (95% confidence). Tierney (priv. comm.) attrib for CMIP [53] (Coupled Model Intercomparison Project) collabora
utes the difference between the two studies to the broader time tion with PMIP [54] (Paleoclimate Modelling Intercomparison
interval of the former study, and concludes that peak LGM cooling Project); a study of LGM surface forcing could aid GCM develop
was near 7� C. ment and assessment of climate sensitivity. Sherwood et al. [21]
Seltzer et al. [51] use the temperature-dependent solubility of review studies of LGM ice sheet forcing and settle on 3.2 ± 0.7 W/
dissolved noble gases in ancient groundwater to show that land m2, the same as IPCC AR4 [55]. However, some GCMs yield effica
areas between 45� S and 35� N cooled 5.8 ± 0.6� C in the LGM. This cies as low as �0.75 [56] or even �0.5 [57], likely due to cloud
cooling is consistent with 1 km lowering of alpine snowlines shielding. We found [7] a forcing of −0.9 W/m2 for LGM vegeta

found by Rind and Peteet [24]. Land response to a forcing exceeds tion by using the Koppen [58] scheme to relate vegetation to local
ocean response, but polar amplification makes the global re climate, but we thought the model effect was exaggerated as
sponse as large as the low latitude land response in GCM simula real-world forests tends to shake off snow albedo effects. Kohler
tions with fixed ice sheets (Supplementary Material Fig. S3). et al. [59] estimate a continental shelf forcing of −0.67 W/m2.
When ice sheet growth is added, cooling amplification at mid Based on an earlier study [60] (hereafter Target CO2), our estimate
and high latitudes is greater [7], making 5.8� C cooling of low lati of LGM-Holocene surface forcing is 3.5 ± 1 W/m2. Thus, LGM (18–
tude land consistent with global cooling of �7� C. 21 kyBP) cooling of 7� C relative to mid-Holocene (7 kyBP), GHG
LGM CO2, CH4 and N2O amounts are known accurately with forcing of 2.25 W/m2, and surface forcing of 3.5 W/m2 yield an
the exception of N2O in the PGM when N2O reactions with dust in initial ECS estimate 7/(2.25 þ 3.5) ¼ 1.22� C per W/m2. We discuss
the ice core corrupt the data. We take PGM N2O as the mean of uncertainties in Equilibrium climate sensitivity section.
the smallest reported PGM amount and the LGM amount; poten PGM-Eemian global warming provides a second assessment of
tial error in the N2O forcing is �0.01 W/m2. We calculate CO2, ECS, one that avoids concern about human influence. PGM-
CH4, and N2O forcings using Equation (4) and formulae for each Eemian GHG forcing is 2.3 W/m2. We estimate surface albedo
gas in Supplementary Material for the periods shown by green forcing as 0.3 W/m2 less than in the LGM because sea level was
bars in Fig. 3. The Eemian period avoids early CO2 and tempera about 10 m higher during the PGM [61]. North American and
ture spikes, assuring that Earth was in energy balance. Between Eurasian ice sheet sizes differed between the LGM and PGM [62],
the LGM (19–21 kyBP) and Holocene (6.5–7.5 kyBP), GHG forcing but division of mass between them has little effect on the net
increased 2.25 W/m2 with 77% from CO2. Between the PGM and forcing (Supplementary Fig. S4 [60]). Thus, our central estimate
Eemian, GHG forcing increased 2.30 W/m2 with 79% from CO2. of PGM-Eemian forcing is 5.5 W/m2. Eemian temperature reached
Glacial-interglacial aerosol changes are not included as a forc about þ1� C warmer than the Holocene [63], based on Eemian
ing. Natural aerosol changes, like clouds, are fast feedbacks. SSTs of þ0.5 ± 0.3� C relative to 1870–1889 [64], or þ0.65 ± 0.3� C
Indeed, aerosols and clouds form a continuum and distinction is SST and þ1� C global (land plus ocean) relative to 1880–1920.
arbitrary as humidity approaches 100%. There are many aerosol However, the PGM was probably warmer than the LGM; it was
types, including VOCs (volatile organic compounds) produced by warmer at Dome C (Fig. 2), but cooler at Dronning Maud Land
trees, sea salt produced by wind and waves, black and organic [65]. Based on deep ocean temperatures (Cenozoic Era section),
carbon produced by forest and grass fires, dust produced by wind we estimate PGM-Eemian warming as 0.5� C greater than LGM-
and drought, and marine biologic dimethyl sulfide and its sec Holocene warming, that is 7.5� C. The resulting ECS is 7.5/5.5 ¼
ondary aerosol products, all varying geographically and in re 1.36� C per W/m2. Although PGM temperature lacks quantifica
sponse to climate change. We do not know, or need to know, tion comparable to that of Seltzer et al. [51] and Tierney et al. [49]
natural aerosol properties in prior eras because their changes are for the LGM, the PGM-Eemian warming provides support for the
feedbacks included in the climate response. However, human- high ECS inferred from LGM-Holocene warming.
made aerosols are a climate forcing (an imposed perturbation of We conclude that ECS for climate in the Holocene-LGM range
Earth’s energy balance). Humans may have begun to affect gases is 1.2� C ± 0.3� C per W/m2, where the uncertainty is the 95% confi
and aerosols in the latter Holocene (Aerosols section), but we dence range. The uncertainty estimate is inherently subjective,
minimize that issue by using the 6.5–7.5 kyBP window to evaluate as it depends mainly on the ice age surface albedo forcing. The
climate sensitivity. GHG forcing and glacial-interglacial temperature change are
Earth’s surface change is the other forcing needed to evaluate well-defined, but the efficacy of ice age surface forcing varies
ECS: (1) change of surface albedo (reflectivity) and topography by among GCMs. This variability is likely related to cloud shielding
5 300 700 300

Temperature (ºC)
Temperature CH4
CO2 (ppm)
CO2 N2 O
N2O (ppb)
CH4 (ppb)
600
0
250 250
500
−5
400
200 200
−10 300
160 140 120 100 80 60 40 20 0 160 140 120 100 80 60 40 20 0
Time (kyBP) Time (kyBP)
Figure 3. Dome C temperature (Jouzel et al. [40]) and multi-ice core GHG amounts (Schilt et al. [47]). Green bars (1–5, 6.5–7.5, 18–21, 120–126,
137–144 kyBP) are periods of calculations.
of surface albedo, which reaffirms the need for a focus on precise response time of 25 years, while several simplified climate mod
cloud observations and modeling. els referenced in our Ewing paper had even faster responses. The
longer response time of our climate model was largely a result of
State dependence of climate sensitivity high climate sensitivity—our model had an ECS of 4� C for 2 �
ECS based on glacial-interglacial climate is an average for global CO2 while the Bryan and Manabe model had an ECS of 2� C.
temperatures −7� C to þ1� C relative to the Holocene and in gen The physics is straightforward. If the delay were a result of a
eral differs for other climate states because water vapor, aerosol- fixed source of thermal inertia, say the ocean’s well-mixed upper
cloud and sea ice feedbacks depend on the initial climate. layer, response time would increase linearly with ECS because
However, ECS is rather flat between today’s climate and warmer most climate feedbacks come into play in response to tempera
climate, based on a study [66] covering a range of 15 CO2 dou ture change driven by the adjusted forcing, not in direct response
blings using an efficient GCM developed by Gary Russell [67]. to the forcing. Thus, a model with ECS of 4� C takes twice as long
Toward colder climate, ice-snow albedo feedback increases non to reach full response as a model with ECS of 2� C, if the mixed
linearly, reaching snowball Earth conditions—with snow and ice layer provides the only heat capacity. However, while the mixed

on land reaching sea level in the tropics—when CO2 declines to a layer is warming, there is exchange of water with the deeper
quarter to an eighth of its 1950 abundance (Fig. 7 of the study) ocean, which slows the mixed layer warming. The longer re
[66]. Snowball Earth occurred several times in Earth’s history, sponse time with high ECS allows more of the ocean to come into
most recently about 600 million years ago [68] when the Sun was play. If mixing into the deeper ocean is approximated as diffu
6% dimmer [69] than today, a forcing of about –12 W/m2. Toward sive, surface temperature response time is proportional to the
warmer climate, the water vapor feedback increases as the tropo square of climate sensitivity [74].
pause rises [70], the tropopause cold trap disappearing at 32 � Slow climate response accentuates need for the ‘anticipation’
CO2 (Fig. 7) [66]. However, for the range of ECS of practical inter that E.E. David, Jr spoke about. If ECS is 4.8� C (1.2� C per W/m2),
est—say from half preindustrial CO2 to 4 � CO2 —state depen more warming is in the pipeline than widely assumed. GHG forc
dence of ECS is small compared to state dependence of ESS. ing today already exceeds 4 W/m2. Aerosols reduce the net forc
Earth system sensitivity (ESS) includes amplifying feedbacks ing to about 3 W/m2, based on IPCC estimates (Aerosols section),
of GHGs and ice sheets [71]. When we consider CO2 change as a but warming still in the pipeline for 3 W/m2 forcing is 2.4� C, ex
known forcing, other GHGs provide a feedback that is smaller ceeding warming realized to date (1.2� C). Slow feedbacks increase
than the ice sheet feedback, but not negligible. Ice core data on the equilibrium response even further (Summary section). Large
GHG amounts show that non-CO2 GHGs—including O3 and warmings can be avoided via a reasoned policy response, but def
stratospheric H2O produced by changing CH4—provide about inition of effective policies will be aided by an understanding of
20% of the total GHG forcing, not only on average for the full climate response time.
glacial-interglacial change, but as a function of global tempera
ture right up to þ1� C global temperature relative to the Holocene Temperature response function
(Supplementary Fig. S5). Atmospheric chemistry modeling sug In the Bjerknes lecture [75] at the 2008 American Geophysical
gests that non-CO2 GHG amplification of CO2 forcing by about a Union meeting, JEH argued that the ocean in many5 GCMs had
quarter continues into warmer climate states [72]. Thus, for cli excessive mixing, and he suggested that GCM groups all report
mate change in the Cenozoic era, we approximate non-CO2 GHG the response function of their models—the global temperature
forcing by increasing the CO2 forcing by one-quarter. change versus time in response to instant CO2 doubling with the
Ice sheet feedback, in contrast to non-CO2 GHG feedback, is model run long enough to approach equilibrium. The response
highly nonlinear. Preindustrial climate was at most a few halv function characterizes a climate model and enables a rapid esti
ings of CO2 from runaway snowball Earth and LGM climate was mate of the global mean surface temperature change in response
even closer to that climate state. The ice sheet feedback is re to any climate forcing scenario:
duced as Earth heads toward warmer climate today because al
ð
ready two-thirds of LGM ice has been lost. Yet remaining ice on � �
TG ðtÞ ¼ ½dTG ðtÞ=dt� dt ¼ S k � RðtÞ dFe =dt dt: (5)
Antarctica and Greenland constitutes a powerful feedback,
which humanity is about to bring into play. We can illuminate
that feedback and the climate path Earth is now on by examining TG is the Green’s function estimate of global temperature at time
data on the Cenozoic era—which includes CO2 levels comparable t, k (� C per W/m2) the model’s equilibrium sensitivity, R the di
to today’s amount—but first we must consider climate re mensionless temperature response function (% of equilibrium re
sponse time. sponse), and dFe the forcing change per unit time, dt. Integration
over time begins when Earth is in near energy balance, e.g. in pre
industrial time. The response function yields an accurate esti
Climate response time mate of global temperature change for a forcing that does not
In this section we define response functions for global tempera cause reorganization of ocean circulation. Accuracy of this ap
ture and Earth’s energy imbalance that help reveal the physics of proximation for temperature for one climate model is shown in
climate change. Cloud feedbacks amplify climate sensitivity and Chart 15 in the Bjerknes presentation and wider applicability has
thus increase eventual heat uptake by the ocean, but cloud feed been demonstrated [76].
backs also have the potential to buffer the rate at which the We study ocean mixing effects by comparing two GCMs: GISS
ocean takes up heat, thus increasing climate response time. (2014) [77] and GISS (2020) [33], both models6 described by Kelley
Climate response time was surprisingly long in our climate et al. [32].Ocean mixing is improved in GISS (2020) by use of a
simulations [7] for the 1982 Ewing Symposium. The e-folding high-order advection scheme [78], finer upper-ocean vertical res
time—the time for surface temperature to reach 63% of its equi olution (40 layers), updated mesoscale eddy parameterization,
librium response—was about a century. The only published and correction of errors in the ocean modeling code [32]. The
atmosphere-ocean GCM—that of Bryan and Manabe [73]—had a GISS (2020) model has improved variability, including the
Madden-Julian Oscillation (MJO), El Nino Southern Oscillation stored in warming continents and heat used in net ice melt [80].
(ENSO) and Pacific Decadal Oscillation (PDO), but the spectrum Heat storage in air adds a small amount. Radiation balance mea
of ENSO-like variability is unrealistic and its amplitude is exces sured from Earth-orbiting satellites cannot by itself define the
sive, as shown by the magnitude of oscillations in Fig. 4a. Ocean absolute imbalance, but, when anchored to an in situ EEI value
mixing in GISS (2020) may still be excessive in the North Atlantic, for a sufficient interval (e.g. 10 years), satellite Earth radiation
where the model’s simulated penetration of CFCs is greater than budget observations [81] provide invaluable EEI data on finer
observed [79]. temporal and spatial scales than the in situ data.
Despite reduced ocean mixing, the GISS (2020) model surface After a step-function forcing is imposed, EEI and global sur
temperature response is no faster than in the GISS (2014) model face temperature must each approach a new equilibrium, but EEI
(Fig. 4b): it takes 100 years to reach within 1/e of the equilibrium does so more rapidly, especially for the GISS (2020) model (Fig. 5).
response. Slow response is partly explained by the larger ECS of EEI in GISS (2020) needs only a decade to reach within 1/e of full
the GISS (2020) model, which is 3.5� C versus 2.7� C for the GISS response (Fig. 5b), but global surface temperature requires a cen
(2014) model, but something more is going on in the newer tury (Fig. 4b). Rapid decline of EEI—to half the forcing in 5 years

model, as exposed by the response function of Earth’s en (Fig. 5a)—has practical implications. First, EEI defines the rate
ergy imbalance. heat is pumped into the ocean, so if EEI is reduced, ocean warm
ing is slowed. Second, rapid EEI decline implies that it is wrong to
Earth’s energy imbalance (EEI) assume that global warming can be stopped by a reduction of cli
When a forcing perturbs Earth’s energy balance, the imbalance mate forcing by the amount of EEI. Instead, the required reduc
drives warming or cooling to restore balance. Observed EEI is tion of forcing is larger than EEI. The difficulty in finding
now of order þ1 W/m2 (more energy coming in than going out) additional reduction in climate forcing of even a few tenths of a
[80]. High accuracy of EEI is obtained by tracking ocean warm W/m2 is substantial [63]. Calculations that help quantify this
ing—the main repository for excess energy—and adding heat matter are discussed in Supplementary Material section SM8.
(a) Global Temperature Change (ºC) (b) Temperature Response Function (%)
4 100
GISS (2014) GISS (2014)

GISS (2020) 80 GISS (2020)
3
1−1/e
60
2
40
1
20
0 0
1 10 100 1000 1 10 100 1000
Year Year
Figure 4. (a) Global mean surface temperature response to instant CO2 doubling and (b) normalized response function (percent of final change). Thick
lines in Figs 4 and 5 are smoothed (yr1 no smoothing; yr2 3-yr mean; yr3–12 5-yr mean, yr13–300 25-yr mean; yr301–5000 101-yr mean).
(a) Earth’s Energy Imbalance (W/m2) (b) EEI Response Function (%)
4 100
GISS (2014)
3 GISS (2020) 80
1−1/e
2 60
1 40
0 20
GISS (2014)
GISS (2020)
−1 0
1 10 100 1000 1 10 100 1000
Year Year
Figure 5. (a) Earth’s energy imbalance (EEI) for 2 � CO2, and (b) EEI normalized response function.
What is the physics behind the fast response of EEI? The 2 � states is hampered by uncertainty in the forcings that maintained
CO2 forcing and initial EEI are both nominally 4 W/m2. In the the climate, as proxy measures of CO2 have large uncertainty.
GISS (2014) model, the decline of EEI averaged over the first year Theory informs us that CO2 is the principal control knob on
is 0.5 W/m2 (Fig. 5a), a moderate decline that might be largely global temperature [87]. Climate of the past 800 000 years dem
caused by warming continents and thus increased heat radiation onstrates (Fig. 2) the tight control. Our aim here is to extract
to space. In contrast, EEI declines 1.3 W/m2 in the GISS (2020) Cenozoic surface temperature history from the deep ocean oxy
model (Fig. 5a). Such a huge, immediate decline of EEI implies ex gen isotope d18O and infer Cenozoic CO2 history. Oxygen isotope
istence of an ultrafast climate feedback. Climate feedbacks are data has high temporal resolution for the entire Cenozoic, which
the heart of climate change and warrant discussion. aids understanding of Cenozoic climate change and resulting
implications for future climate. Our CO2 analysis is a comple
Slow, fast and ultrafast feedbacks ment to proxy CO2 measurements. Despite progress in estimating
Charney et al. [4] described climate feedbacks without discussing CO2 via carbon isotopes in alkenones and boron isotopes in
time scales. At the 1982 Ewing Symposium, water vapor, clouds planktic foraminifera [88], there is wide scatter among results

and sea ice were described as ‘fast’ feedbacks [7] presumed to and fossil plant stomata suggest smaller CO2 amounts [89].
change promptly in response to global temperature change, as
opposed to ‘slow’ feedbacks or specified boundary conditions Deep ocean temperature and sea level from d18O
such as ice sheet size, vegetation cover, and atmospheric CO2 Glacial-interglacial CO2 oscillations (Fig. 2) involve exchange of
amount, although it was noted that some specified boundary carbon among surface carbon reservoirs: the ocean, atmosphere,
conditions, e.g. vegetation, in reality may be capable of relatively soil and biosphere. Total CO2 in the reservoirs also can vary,
rapid change [7]. mainly on longer time scales, as carbon is exchanged with the
The immediate EEI response (Fig. 5a) implies a third feedback solid Earth. CO2 then becomes a primary agent of long-term cli
time scale: ultrafast. Ultrafast feedbacks are not a new concept. mate change, leaving orbital effects as ‘noise’ on larger climate
When CO2 is doubled, the added infrared opacity causes the swings. Oxygen isotopic composition of benthic (deep ocean
stratosphere to cool. Instant EEI upon CO2 doubling is only Fi ¼ dwelling) foraminifera shells provides a starting point for analy
þ2.5 W/m2, but stratospheric cooling quickly increases EEI to þ4 sis of Cenozoic temperature. Figure 6 includes the recent high-
W/m2 [82]. All models calculate a similar radiative effect, so it is resolution record of Westerhold et al. [90] and data of Zachos
useful to define an adjusted forcing, Fa, which is superior to Fi as et al. [44] that have been used for many studies in the past quar
a measure of climate forcing. In contrast, if cloud change—the ter century. When Earth has negligible ice sheets, d18O (18O
likely cause of the present ultrafast change—is lumped into the amount relative to a standard), provides an estimate of deep
adjusted forcing, each climate model has its own forcing, losing ocean temperature (right scale in Fig. 6) [44].
the merit of a common forcing.
Kamae et al. [83] review rapid cloud adjustment distinct from Tdo ð� CÞ ¼ 4 d18 O þ 12: (6)
surface temperature-mediated change. Clouds respond to radia
tive forcing, e.g. via effects on cloud particle phase, cloud cover, This equation is used for the early Cenozoic, up to the large-scale
cloud albedo and precipitation [84]. The GISS (2020) model alters glaciation of Antarctica at �34 MyBP (Oi-1in Fig. 6). At larger d18O
glaciation in stratiform mixed-phase clouds, which increases (colder climate), lighter 16O evaporates preferentially from the
supercooled water in stratus clouds, especially over the Southern ocean and accumulates in ice sheets. In Zachos data, d18O
Ocean [Fig. 1 in the GCM description [32]]. The portion of super increases by 3 between Oi-1 and the LGM. Half of this d18O
cooled cloud water drops goes from too little in GISS (2014) to too change is due to the 6� C change of deep ocean temperature be
much in GISS (2020). Neither model simulates well stratocumu tween Oi-1 (5� C) and the LGM (–1� C) [92]. The other 1.5 of d18O
lus clouds, yet the models help expose real-world physics that change is presumed to be due to the �180 m sea level (SL) change
affects climate sensitivity and climate response time. Several between ice-free Earth and the LGM, with �60 m from Antarctic
models in CMIP6 comparisons find high ECS [84]. For the sake of ice and 120 m from Northern Hemisphere ice. Thus, as an ap
revealing the physics, it would be useful if the models defined proximation to extract both SL and Tdo from d18O, Hansen et al.
their temperature and EEI response functions. Model runs of
[66] assumed that SL rose linearly by 60 m as d18O increased from
even a decade can define the important part of Figs 4a and 5a.
1.75 to 3.25 and linearly by 120 m as d18O increased from 3.25
Many short (e.g. 2-year) 2 � CO2 climate simulations with each
to 4.75.
run beginning at a different point in the model’s control run, can
The Zachos (Z) and Westerhold (W) d18O time series differ
define cloud changes to an arbitrary accuracy.
(Fig. 6) mainly because of different sites of the sediment cores
and the way multiple sites are stacked to obtain a time series for
Cenozoic era the full Cenozoic. For example, mid-Holocene (6–8 kyBP) values
of d18O in the Z and W data sets are d18OZH ¼ 3.32 and d18OW H ¼
In this section, we use ocean sediment core data to explore climate
3.88. Thus, the Z and W d18O time series require separate equa
change in the past 66 million years. This allows us to study warmer
tions for sea level (SL) and deep ocean temperature (Tdo) [66]:
climates that are relevant to human-made climate forcing.
High equilibrium climate sensitivity that we have inferred,
SLZ ðmÞ ¼ 60 38:2 ðd18 O 1:75Þ (7)
ECS ¼ 1.2� C ± 0.3� C per W/m2, may affect interpretation of
18
warmer climates. GCMs have difficulty in producing Pliocene ðd O < 3:32; maximum SL ¼ þ60 mÞ;
warmth [85], especially in the Arctic, without large—probably un
realistic—CO2 amounts. In addition, a coupled GCM/ice sheet
model needs 700–840 ppm CO2 for transition between glaciated SLW ðmÞ ¼ 60 25:2 ðd18 O 1:5Þ (8)
and unglaciated Antarctica [86]. Understanding of these climate ðd18 O < 3:88; maximum SL ¼ þ60 mÞ;

Figure 6. Global deep ocean d18O. Black line: Westerhold et al. [90] data in 5 kyr bins until 34 MyBP and subsequently 2 kyr bins. Green line: Zachos et al.
[44] data at 1 Myr resolution. Lower left: velocity [91] of Indian tectonic plate. PETM ¼ Paleocene Eocene Thermal Maximum; EECO ¼ Early Eocene
Climatic Optimum; Oi-1 marks the transition to glaciated Antarctica; MCO ¼ Miocene Climatic Optimum; NAIP ¼ North Atlantic Igneous Province.
120ðd18 O 3:32Þ ðd18 O > 3:32Þ; We expect Tdo change, which derives from sea surface tempera
SLZ ðmÞ ¼ (9)
1:58 ture (SST) at high latitudes where deepwater forms, to approximate
TS change when Tdo is not near the freezing point. Global SST
change understates global TS (land plus ocean) change because
120ðd18 O 3:88Þ ðd18 O > 3:88Þ;
SLW ðmÞ ¼ (10) land temperature response to a forcing exceeds SST response [95],
1:42
e.g. the equilibrium global SST response of the GISS (2020) GCM to 2
where 1.75 and 1.5 are d18O midpoints at the Oi-1 transition for � CO2 is 70.6% of the global (land plus ocean) response. However,
the Z and W data sets. Equations (9) and (10) are based on polar amplification of the SST response tends to compensate for
d18OZLGM ¼ 4.9 and d18OW SST undershoot of global TS change. Compensation is nearly exact
LGM ¼ 5.3 with SL ¼ 0 today. Tdo equations
are based on specified Holocene and LGM Tdo of 1� C [93] and at latitudes of North Atlantic deepwater formation for 2 � CO2 cli
−1� C [92], respectively. Coefficients in the Tdo equations are cal mate change in the GISS (2020) climate model (Fig. 7a), but
culated as shown by the Equation (12) example. Southern Hemisphere polar amplification does not fully cover the
60–75� S latitudes where Antarctic bottom water forms.
TZdo ð� CÞ ¼ 5 2:55 ðd18 O 1:75Þ ð1:75 < d18 O < 3:32Þ; (11) As Tdo nears the freezing point, ice forms, adhering to the
Antarctic continent, extending today to a depth of about 2 km,
and also forming floating ice shelves. From the Holocene toward
TZdo ð� CÞ ¼ 1 2 ðd18 O 3:32Þ=ð4:9 3:32Þ colder climate, the effect on temperature change is large: TS
¼1 1:27ðd18 O 3:32Þ ð3:32 < d18 OÞ; (12) declines 7� C between the Holocene and LGM, but Tdo declines only
2� C (from 1� C to –1� C). From the Holocene toward hotter climate,
we expect a smaller effect that we quantify by first neglecting the
TW ð� Þ
do C ¼ 6 2:10 ðd18 O 1:5Þ ð1:5 < d18 O < 3:88Þ; (13) effect and finding how far we underestimate EECO temperature.
Thus, as an initial approximation we assume DTS ¼ DTdo:
TW ð� Þ
do C ¼ 1 1:41 ðd18 O 3:88Þ ð3:88 < d18 OÞ: (14)
TS � Tdo TdoH þ 14� C ¼ Tdo þ 13� C; ðd18 O < d18 OH Þ (15)
18
Zachos and Westerhold d O, SL and Tdo for the full Cenozoic,
where we take Holocene TS as 14� C and TdoH as 1� C. In this initial
Pleistocene, and the past 800 000 years are graphed in
approximation, we interpolate linearly for climate colder than
Supplementary Material and sea level is compared to data of
the Holocene, the LGM being �7� C cooler than the Holocene:
Rohling et al. [94]. We focus on the finer resolution W data.
Differences between the W and Z data and interpretation of
those differences are discussed in Paleocene Eocene Thermal TS ¼ 14� C 7� C � ðd18 O d18 OH Þ (16)
Maximum section. =ðd18 OLGM d18 OH Þ ðd18 O > d18 OH Þ
Cenozoic TS Resulting EECO (Early Eocene Climatic Optimum) TS is �27� C

In this section we combine the rich detail in Tdo provided by (Fig. 8a). As expected, this initial approximation undershoots EECO
benthic d18O with constraints on the range of Cenozoic TS from TS, which Zhu et al. [96] infer to be 29� C from a proxy-constrained
surface proxies to produce an estimated history of Cenozoic TS. full-field analysis using a GCM to account for the pattern of
(a) (b)

Figure 7. (a) Ratio of DSST (latitude) to global TS change for all ocean and the Atlantic Ocean, based on equilibrium response (years 4001–4500) in
2 � CO2 simulations of GISS (2020) model. (b) DT, the amount by which TS change exceeds Tdo change, based on an exponential fit to the two data
points provided by the Holocene and LGM (see text).
(a) Surface Temperature (initial linear approx.) (b) Surface Temperature (from equation 16)
30 30
Past 67 Million Years Past 67 Million Years
25 25
Ts (°C)
Zhu (2019) Zhu (2019)

20 20
15 15
10 10
65 60 55 50 45 40 35 30 25 20 15 10 5 0 65 60 55 50 45 40 35 30 25 20 15 10 5 0
18 18
16 16
14 14
Ts (°C)
12 12
10 10
8 Past 5.33 Million Years 8 Past 5.33 Million Years
6 6
5 4 3 2 1 0 5 4 3 2 1 0
18 18
16 Past 800 Thousand Years 16 Past 800 Thousand Years
14 14
Ts (°C)
12 12
10 10
8 8
6 6
.8 .7 .6 .5 .4 .3 .2 .1 0. .8 .7 .6 .5 .4 .3 .2 .1 0.
Time (MyBP) Time (MyBP)
Figure 8. Cenozoic temperature based on linear (Equations 15 and 16) and nonlinear (Equation 17) analyses. Antarctic Dome C data [40] (red) relative to
last 1000 years are multiplied by 0.6 to account for polar amplification and 14� C is added for absolute scale.
temperature change. Moderate undershoot (DT ¼ 2� C) of EECO TS The result is a consistent analysis of global TS for the entire
is consistent with expectation that global warming of a few Cenozoic (Fig. 8b). Oxygen isotope d18O of deep ocean forami
degrees would remove Antarctic ice shelves and allow polar ampli nifera reproduces glacial-interglacial temperature change well;
fication to fully cover regions of deepwater formation. Moreover, more detailed agreement is not expected as Antarctic ice core
DT of 2� C at the Holocene and 5� C more between the Holocene data are for a location that moves, especially in altitude. Our
and LGM are fit well by an exponential function between Antarctic interest is in warmer global climate and its relevance to up
glaciation and the LGM, as needed for DT to asymptote at the coming human-caused climate change. For that purpose, we
freezing point (Fig. 7b). Thus, we take TS as want to know the forcing that drove Cenozoic climate
change. With the assumption that non-CO2 GHG forcings pro
TS ¼ Tdo DT þ 15� C ¼ Tdo 0:35ðe0:8X 1Þ þ 15� C; (17) vide 20% of the total GHG forcing, it is not difficult to infer
the CO2 abundance required to cause the Cenozoic tempera
where X ¼ d18O − d18OOi-1 and TS is normalized to 14� C in ture history in Fig. 8b. Considering the large disagreement
the Holocene. among proxy CO2 measures, this indirect measure of
CO2 via global TS may provide the most accurate Cenozoic For climate warmer than the Holocene up to Oi-1, i.e. for
CO2 history. d18OOi-1 < d18O < d18OH,
Cenozoic CO2 " #

d18 OH d18 O
We obtain the CO2 history required to yield the Cenozoic TS his DFðtÞ ¼ 1:25 � DFCO2 ðtÞþDFSol ðtÞ þ FIceH � :
d18 OH d18 OOi 1
tory from the relation
(22)
�
DFðtÞ ¼ ½TS ðtÞ 14 C�=ECS; (18)
FIceH, the (Antarctic plus Greenland) ice sheet forcing between the
Holocene and Oi-1, is estimated to be 2 W/m2 (Supplementary Fig.
where DF(t) (0 at 7 kyBP) includes changing solar irradiance and
S4, Target CO2). For climate warmer than Oi-1
amplification of CO2 forcing by non-CO2 GHGs and ice sheets.
The GHG amplification factor is taken as 1.25 throughout the
DFðtÞ ¼ 1:25 � ½DFCO2 ðtÞ þ DFSol ðtÞ þ DFIceH � (23)
Cenozoic (State dependence of climate sensitivity section).

The amplification applies to solar forcing as well as CO2 forcing All quantities are known except DFCO2(t), which is thus defined.
because it is caused by temperature change, not by CO2. Solar Cenozoic CO2 (t) for specified ECS is obtained from TS(t) using the
irradiance is increasing 10% per billion years [69]; thus solar forc CO2 radiative forcing equation (Table 1, Supplementary
ing (240 W/m2 today) increases 2.4 W/m2 per 100 million Material). Resulting CO2 (Fig. 9) is about 1,200 ppm at the EECO,
years. Thus, 450 ppm at Oi-1, and 325 ppm in the Pliocene for ECS ¼ 1.2� C per
� � W/m2. For ECS ¼ 1� C—about as low as we believe plausible—
DFðtÞ ¼ 1:25 � DFCO2 ðtÞ þ DFSol ðtÞ � AS : ðd18 O > d18 OH Þ (19) Pliocene CO2 is near 350 ppm, rising only to �500 ppm at Oi-1
and �1500 ppm at EECO.
AS, surface albedo amplification, is smaller in moving from the Assumed Holocene CO2 amount is also a minor factor. We
Holocene to warmer climate—when the main effect is shrinking tested two cases: 260 and 278 ppm (Fig. 9). These were imple
of Antarctic ice—than toward colder climate. For d18O > d18OH, mented as the CO2 values at 7 kyBP, but Holocene-mean values
we take AS as its average value over the period from the are similar—a few ppm less than CO2 at 7 kyBP. Holocene ¼ 278
Holocene to the LGM: ppm increases CO2 about 20 ppm between today and Oi-1, and
about 50 ppm at the EECO. However, Holocene CO2 278 ppm
FIce þFGHG 3:5þ2:25 �
AS ¼ ¼ ¼ 2:55: d18 O > d18 OH (20) causes the amplitude of inferred glacial-interglacial CO2 oscilla
FGHG 2:25
tions to be less than reality (Fig. 9b), providing support for the
Holocene 260 ppm level and for the interpretation that high late-
Thus, for climate colder than the Holocene,
Holocene CO2 was due to human influence. Proxy measures of
� � Cenozoic CO2 yield a notoriously large range. A recent review [88]
DFðtÞ ¼ 3:19 � DFCO2 ðtÞ þ DFSol ðtÞ : ðd18 O > d18 OH Þ (21)
constructs a CO2 history with Loess-smoothed CO2 �700–1100
(a) Holocene CO2 = 260 ppm (b) Holocene CO 2 = 278 ppm

2000 2000
1500 Past 67 Million Years 1500 Past 67 Million Years
1000 1000
500 450 ppm 500 450 ppm

ECS = 1.0 ECS = 1.0
ECS = 1.4 300 ppm ECS = 1.4 300 ppm
200 200
65 60 55 50 45 40 35 30 25 20 15 10 5 0 65 60 55 50 45 40 35 30 25 20 15 10 5 0
350 350
300 300
250 250
200 Past 5.33 Million Years ECS = 1.0 200 Past 5.33 Million Years ECS = 1.0
ECS = 1.4 ECS = 1.4
5 4 3 2 1 0 5 4 3 2 1 0
350 350
Past 800 Thousand Years Past 800 Thousand Years
300 300
250 250
200 200
ECS = 1 ECS = 1
ECS = 1.4 ECS = 1.4
150 150
.8 .7 .6 .5 .4 .3 .2 .1 0. .8 .7 .6 .5 .4 .3 .2 .1 0.
Figure 9. Cenozoic CO2 estimated from d18O of Westerhold et al. (see text). Black lines are for ECS ¼ 1.2� C per W/m2; red and green curves (ECS ¼ 1.0 and
1.4� C per W/m2) are 1 My smoothed. Blue curves (last 800 000 years) are Antarctica ice core data [41].
ppm at Oi-1. That high Oi-1 CO2 amount is not plausible without the Indian plate, but burial of organic matter and increased
overthrowing the concept that global temperature is a response weathering due to exposure of fresh rock by Himalayan uplift
to climate forcings. More generally, we conclude that actual CO2 [105] may contribute to CO2 drawdown. Quantitative under
during the Cenozoic was near the low end of the range of proxy standing of these processes is limited [106], e.g. weathering is
measurements. both a source and sink of CO2 [107].
This picture for the broad sweep of Cenozoic CO2 is consistent
Interpretation of Cenozoic TS and CO2 with current understanding of the long-term carbon cycle [108],
In this section we consider Cenozoic TS and CO2 histories, which but relative contributions of metamorphism [106] and volcanism
are rich in insights about climate change with implications for [109] are uncertain. Also, emissions from rift-induced Large
future climate. Igneous Provinces (LIPs) [110, 111] contribute to long-term
In Target CO2 [60] and elsewhere [98] we argue that the broad change of atmospheric CO2, with two cases prominent in Fig. 6.
sweep of Cenozoic temperature is a result of plate tectonic (popu The Columbia River Flood Basalt at ca. 17–15 MyBP was a princi
larly ‘continental drift’) effects on CO2. Solid Earth sources and pal cause of the Miocene Climatic Optimum [112], but the pro

sinks of CO2 are not balanced at any given time. CO2 is removed cesses are poorly understood [113]. A more dramatic event
from surface reservoirs by: (1) chemical weathering of rocks with occurred as Greenland separated from Europe, causing a rift in
deposition of carbonates on the ocean floor, and (2) burial of or the sea floor; flood basalt covered more than a million square
ganic matter [99, 100]. CO2 returns via metamorphism and volca kilometers with magma volume 6–7 million cubic kilometers
nic outgassing at locations where oceanic crust is subducted [111]—the North Atlantic Igneous Province (NAIP). Flood basalt
beneath moving continental plates. The interpretation in Target volcanism occurred during 60.5–54.5 MyBP, but at 56.1 ± 0.5
CO2 was that the main Cenozoic source of CO2 was associated MyBP melt production increased by more than a factor of 10, con
with the Indian plate (Fig. 10), which separated from Pangea in tinued at a high level for about a million years, and then subsided
the Cretaceous [101, 102] and moved through the Tethys (now (Fig. 5 of Storey et al. [114]). The striking Paleocene-Eocene
Indian) Ocean at a rate exceeding 10 cm/year until collision with Thermal Maximum (PETM) d18O spike (Fig. 6) occurs early in this
the Eurasian plate at circa 50 MyBP. Associated CO2 emissions in million-year bump-up of d18O. Svensen et al. [115] proposed that
clude those from formation of the Deccan Traps7 in western the PETM was initiated by the massive flood basalt into carbon-
India (a large igneous province, LIP, formed by repeated deposi rich sedimentary strata. Gutjahr et al. [116] developed an isotope
tion of large-scale flood basalts), the smaller Rajahmundry Traps analysis, concluding that most of PETM carbon emissions were
[103] in eastern India, and metamorphism and vulcanism associ volcanic, with climate-driven carbon feedbacks playing a lesser
ated with the moving Indian plate. The Indian plate slowed circa role. Yet other evidence [117], while consistent with volcanism as
60 Mya (inset, Fig. 6) before resuming high speed [91], leaving an a trigger for the PETM, suggests that climate feedback—perhaps
indelible signature in the Cenozoic d18O history (Fig. 6) that sup methane hydrate and peat CO2 release—may have caused more
ports our interpretation of the CO2 source. Since the continental than half of the PETM warming. Berndt et al. [118] describe exten
collision, subduction and CO2 emissions continue at a diminish sive shallow-water vents that likely released CH4 as well as CO2
ing rate as the India plate underthrusts the Asian continent and during the NAIP activity. We discuss PETM warming and CO2 lev
pushes up the Himalayan mountains [104]. We interpret the de els below, but first we must quantify the mechanisms that drove
cline of CO2 over the past 50 million years as, at least in part, a Cenozoic climate change and consider where Earth’s climate was
decline of the metamorphic source from continued subduction of headed before humanity intervened.
Figure 10. Continental configuration 56 MyBP [97]. Continental shelves (light blue) were underwater as little water was locked in ice. The Indian plate
was moving north at about 15 cm per year.
The sum of climate forcings (CO2 and solar) and slow feed most recent large igneous province (LIP) event—the Columbia
backs (ice sheets and non-CO2 GHGs) that maintained EECO River Flood Basalt about 15 million years ago (Fig. 6)—is no longer
warmth was 12.5 W/m2 (Fig. 11). CO2 forcing of 9.1 W/m2 com a factor, and there is no evidence of another impending LIP.
bined with solar forcing of—1.2 W/m2 to yield a total forcing8 8 Snowball conditions are possible, even though the Sun’s bright
W/m2. Slow feedbacks were 4.5 W/m2 forcing (ice albedo ¼ 2 W/ ness is increasing and is now almost 6% greater [69] than it was
m2 and non-CO2 GHGs ¼ 2.5 W/m2). With today’s solar irradi at the last snowball Earth, almost 600 million years ago [68].
ance, human-made GHG forcing required for Earth to return to Runaway snowball likely requires only 1–2 halvings [66] of CO2
EECO warmth is 8 W/m2. Present human-made GHG forcing is from the LGM 180 ppm level, i.e. to 45–90 ppm. Although the
4.6 W/m2 relative to 7 kyBP.9 Equilibrium response to this forcing weathering rate declines in colder climate [119], weathering and
includes the 2 W/m2 ice sheet feedback and 25% amplification (of burial of organic matter continue, so decrease of atmospheric
6.6 W/m2) by non-CO2 GHGs, yielding a total forcing plus slow CO2 could have continued over millions of years, if the source of
feedbacks of 8.25 W/m2. Thus, equilibrium global warming for CO2 from metamorphism and vulcanism continued to decline.
today’s GHGs is 10� C.10 If human-made aerosol forcing is −1.5 W/ Another factor that may have contributed to cooling in the

m2 and remains at that level indefinitely, equilibrium warming Pliocene is uplift and poleward movement of Greenland that ac
for today’s atmosphere is reduced to 8� C. Either 10� C or 8� C celerated about 5 MyBP [120], which likely enhanced glaciation of
dwarfs observed global warming of 1.2� C to date. Most of the Greenland and should be accounted for in simulations of
equilibrium warming for today’s atmosphere has not yet oc Pliocene climate change. We conclude that, in the absence of hu
curred and need not occur (Earth’s energy imbalance section). man activity, Earth may have been headed for snowball Earth
conditions within the next 10 or 20 million years, but the chance
Prospects for another snowball Earth of future snowball Earth is now academic. Human-made GHG
We would be remiss if we did not comment on the precipitous de emissions remove that possibility on any time scale of practical
cline of Earth’s temperature over the last several million years. interest. Instead, GHG emissions are now driving Earth toward
Was Earth falling off the table into another Snowball Earth? much warmer climate.
Global temperature plummeted in the past 50 million years,
with growing, violent, oscillations (Figs 6 and 7). Glacial- Paleocene eocene thermal maximum (PETM)
interglacial average CO2 declined from about 325 ppm to 225 The PETM event provides a benchmark for assessing the poten
ppm in the past five million years in an accelerating decline tial impact of the human-made climate forcing and the time
(Fig. 9a). As CO2 fell to 180 ppm during recent glacial maxima, an scale for natural recovery of the climate system.
ice sheet covered most of Canada and reached midlatitudes in Westerhold [90] data have 10� C deep ocean warming at the
the U.S. Continents in the current supercontinent cycle [101] are PETM (Figs 8 and 12a), which exceeds proxy-derived surface
now dispersed, with movement slowing to 2–3 cm/year. warming. Low latitude SST data have 3–4� C PETM warming [121].
Emissions from the last high-speed high-impact tectonic event— Tierney et al. [122] obtain PETM global surface warming 5.6� C
collision of the Indian plate with Eurasia—are fizzling out. The (5.4–59� C, 95% confidence) via analysis of proxy surface tempera
ture data that accounts for patterns of temperature change.
Zachos [44] data have a deep ocean warming similar to the
proxy-based surface warming. These warming estimates can be
reconciled, but first let’s note the practical importance of
the PETM.
Pre-PETM (56–56.4 MyBP) CO2 is 910 ppm in our analysis for
the most likely ECS (1.2� C per W/m2). Peak PETM CO2 required to
yield the 5.6� C global surface warming estimate of Tierney et al.
[122] is then 1630 ppm if CO2 provides 80% of the GHG forcing,
thus less than a doubling of CO2. (In the unlikely case that CO2
caused 100% of the GHG forcing, required CO2 is 1780, not quite a
doubling.) CO2 amounts for ECS ¼ 1.0 and 1.4� C per W/m2 are
1165 and 760 ppm in the pre-PETM and 2260 and 1270 ppm at
peak PETM, respectively. In all these ECS cases, the CO2 forcing of
Figure 11. Climate forcings and slow feedbacks relative to 7 kyBP from the PETM is less than or about a CO2 doubling. Our assumed 20%
terms in Equations (21–23). contribution by non-CO2 GHGs (amplification factor 1.25,
(a) (b)
40 3700
Surface Temperature (°C) 3000 CO 2 (ppm)
ECS = 1.0°C/(W/m2)
35 2000 ECS = 1.2°C/(W/m2)
ECS = 1.4°C/(W/m2)
30
1000
25 700
56.4 56.2 56.0 55.8 55.6 55.4 56.4 56.2 56.0 55.8 55.6 55.4
Figure 12. Temperature and CO2 implied by Westerhold et al. [90] d18O, if surface warming equaled deep ocean warming. In reality, the unique PETM
event had surface warming �5.6� C, which implies a peak PETM CO2 of about 1630 ppm (see text).
Climate sensitivity (ECS and ESS) section), is nominal; Hopcroft et (Kerguelen Plateau and Maud Rise), which were intended for
al., e.g. estimate a 30% contribution from non-CO2 GHGs [123], study of climate of Antarctica and the Southern Ocean.
thus an amplification factor 1.43. Differences between the W and Z data sets have limited effect
Thus, today’s human-made GHG forcing (4.6 W/m2, growing on our paper, as we apply separate scaling (Equations 7–14) to W
0.5 W/m2 per decade) is already at least comparable to the PETM and Z data to match observations at the LGM, mid-Holocene, and
forcing, although the net human-made forcing including aerosols Oi-1 points. This approach addresses, e.g. the cumulative effect
has probably not reached the PETM forcing. However, there are in combining data splices noted by Zachos in SM9. Further, we
two big differences between the PETM and today. First, there set the EECO global temperature relative to the Holocene and the
were no large ice sheets on Earth in the PETM era. Ice sheets on PETM temperature relative to pre-PETM based on proxy-
Antarctica and Greenland today make Earth system sensitivity constrained, full-field, GCM analyses of Tierney et al. [122] and
(ESS) greater than it was during the PETM. Equilibrium response Zhu et al. [96] Nevertheless, there is much to learn from more
to today’s GHG climate forcing would include deglaciation of precise study of the Cenozoic in general and the PETM
Antarctica and Greenland, sea level rise of 60 m (200 feet), and in particular.

surface albedo forcing (slow feedback) of 2 W/m2. The second dif Policy implications require first an understanding of the role
ference between the PETM and today is the rate of change of the of aerosols in climate change.
climate forcing. Most of today’s climate forcing was introduced
in a century, which is 10 times or more faster than the PETM forc
ing growth. Although a bolide impact [124] has been proposed as
Aerosols
a trigger for the PETM, the issue is the time scale on which the cli The role of aerosols in climate change is uncertain because aero
mate forcing—increased GHGs—occurred. Despite uncertainty in sol properties are not measured well enough to define their cli
the carbon source(s), data and modeling point to duration of a mate forcing. In this section we estimate aerosol climate forcing
millennium or more for PETM emissions [121, 125]. via aerosol effects on Earth’s temperature and Earth’s en
Better understanding of the PETM could inform us on climate ergy imbalance.
feedbacks. Gutjahr et al. [116] argue persuasively that PETM emis Aerosol impact is suggested by the gap between observed
sions were mostly volcanic, yet we know of no other large igne global warming and expected warming due to GHGs based on
ous province that produced such great, temporally-isolated, ECS inferred from paleoclimate (Fig. 13). Expected warming is
emissions. Further, Cenozoic orbitally-driven hyperthermal from Eq. 5 with the normalized response function of the GISS
events [126] testify to large CO2 feedbacks. Northern peatlands (2020) model. Our best estimate for ECS, 1.2� C per W/m2, yields a
today contain more than 1000 Gt carbon [127], much of which gap of 1.5� C between expected and actual warming in 2022.
can be mobilized at PETM warming levels [128]. The double peak Aerosols are the likely cooling source. The other negative forcing
in deep ocean d18O (thus in temperature, cf. Fig. 12, where each discussed by IPCC—surface albedo change—is estimated by IPCC
square is a binning interval of 5000 years) is also found in terres (Chapter 7, Table 7.8) to be –0.12 ± 0.1 W/m2, an order of magni
trial data [129]. Perhaps the sea floor rift occurred in two bursts, tude smaller than aerosol forcing [12]. Thus, for clarity, we focus
or the rift was followed tens of thousands of years later by meth on GHGs and aerosols.
ane hydrate release as a feedback to the ocean warming; much of Absence of global warming over the period 1850–1920
today’s methane hydrate is in stratigraphic deposits hundreds of (Supplementary Fig. S1 of IPCC AR6 WG1 report [12]) is a clue
meters below the sea floor, where millennia may pass before a about aerosol forcing. GHG forcing increased 0.54 W/m2 in 1850–
thermal wave from the surface reaches the deposits [130]. 1920, which causes expected warming 0.3–0.4� C by 1920 for ECS
Feedback emissions, especially from permafrost, seem to be ¼ 1.2� C per W/m2 (Equation 5). Natural forcings—solar irradiance
more chronic than catastrophic, but stabilization of climate may and volcanoes—may contribute to lack of warming, but a persua
require cooling that terminates growth of those feedbacks sive case for the required forcing has not been made. Human-
(Summary section). The PETM provides perhaps the best empiri made aerosols are the likely offset of GHG warming. Such aerosol
cal check on understanding of the atmospheric lifetime of fossil
fuel CO2 [131], but for that purpose we must untangle as well as
possible the time dependence of the PETM CO2 source and feed 2.5
backs. If continuing magma flow or a slow-release feedback is a Observed Global Temperature
substantial portion of PETM CO2, the CO2 lifetime inferred from GHG Warming [ECS = 1.2°C/(W/m2)]
2.0
post-PETM CO2 recovery may be an exaggeration.
GHG Warming (ECS = 1.0°C/(W/m2)]
The PETM draws attention to differences between the
Aerosol Cooling (Faustian Bargain)
1.5
→
Westerhold (W) and Zachos (Z) d18O data. Zachos attributes the
larger PETM response in W data to the shallow (less than 1 km)
depth of the Walvis Ridge core in the Southeast Atlantic that
1.0
anchors the PETM period in the W data (see Supplementary
Material SM9). Given that the PETM was triggered by a rift in the .5
floor of the North Atlantic and massive lava injection, it is not
surprising that ocean temperature was elevated and circulation 0.
disrupted during the PETM. Nunes and Norris [132] conclude that
ocean circulation changed at the start of the PETM with a shift in 1850 1875 1900 1925 1950 1975 2000 2025
location of deep-water formation that delivered warmer waters
Figure 13. Observed global surface temperature (black line) and
to the deep sea, a circulation change that persisted at least
expected GHG warming with two choices for ECS. The blue area is the
40 000 years. With regard to differences in the early Cenozoic, estimated aerosol cooling effect. The temperature peak in the World
Zachos notes (Supplementary Material SM9) a likely bias in the Z War II era is in part an artifact of inhomogeneous ocean data in that
data with a heavy weighting of data from Southern Ocean sites period [63].
cooling is a Faustian bargain [98] because payment in enhanced large sources are not needed to account for Holocene GHG levels.
global warming will come due once we can no longer tolerate the Paleoclimate GHG decreases are slow feedbacks that occur in
air pollution. Ambient air pollution causes millions of deaths per concert with global cooling. However, if global cooling did not oc
year, with particulates most responsible [133, 134]. cur in the past 6000 years, feedbacks did not occur. Earth orbital
parameters 6000 years ago kept the Southern Ocean warm, as
Evidence of aerosol forcing in the Holocene needed to maintain strong overturning ocean circulation [137]
In this section we infer evidence of human-made aerosols in the and minimize carbon sequestration in the deep ocean. Maximum
last half of the Holocene from the absence of global warming. insolation at 60� S was in late-spring (mid-November); since then,
Some proxy-based analyses [135] report cooling in the last half of maximum insolation at 60� S slowly advanced through the year,
the Holocene, but a recent analysis [50] that uses GCMs to over recently reaching mid-summer (mid-January, Fig. 26b of Ice Melt
come spatial and temporal biases in proxy data finds rising global [13]). Maximum insolation from late-spring through mid-
temperature in the first half of the Holocene followed by nearly summer is optimum to warm the Southern Ocean and promote
constant temperature in the last 6000 years until the last few early warm-season ice melt, which reduces surface albedo and

centuries (Fig. 14). Antarctic, deep ocean, and tropical sea surface magnifies regional warming [45].
data all show stable temperature in the last 6000 years GHG forcing of –0.2 W/m2 in 10–6 kyBP (Fig. 15) was exceeded
(Supplementary Fig. S6 of reference [60]). GHG forcing increased by forcing of þ1 W/m2 due to ice sheet shrinkage (Supplementary
0.5 W/m2 during those 6000 years (Fig. 15), yet Earth did not Material in Target CO2 [60]) for a 40 m sea level rise (Fig. 16). Net
warm. Fast feedbacks alone should yield at least þ0.5� C warming 0.8 W/m2 forcing produced expected 1� C global warming (Fig. 14).
and 6000 years is long enough for slow feedbacks to also contrib The mystery is the absence of warming in the past 6000 years.
ute. How can we interpret the absence of warming? Hansen et al. [45] suggested that aerosol cooling offset GHG
Humanity’s growing footprint deserves scrutiny. Ruddiman’s warming. Growing population, agriculture and land clearance
suggestion that deforestation and agriculture began to affect CO2 produced aerosols and CO2; wood was the main fuel for cooking
6500 year ago and rice agriculture began to affect CH4 5000 years and heating. Nonlinear aerosol forcing is largest in a pristine at
ago has been criticized [46] mainly because of the size of pro mosphere, so it is unsurprising that aerosols tended to offset CO2
posed sources. Ruddiman sought sources sufficient to offset warming as civilization developed. Hemispheric differences
declines of CO2 and CH4 in prior interglacial periods, but such could provide a check. GHG forcing is global, while aerosol forc
ing is mainly in the Northern Hemisphere. Global offset implies a
GMST (°C)
1 net negative Northern Hemisphere forcing and positive Southern

0 0 Hemisphere forcing. Thus, data and modeling studies (including
-1 orbital effects) of regional response are warranted but beyond
-2
the scope of this paper.
GMST (°C)
-4 Industrial era aerosols

1850-2019 GMST
Scientific advances often face early resistance from other scien
-6
tists [139]. Examples are the snowball Earth hypothesis [140] and
50
10 95
20 90
30 80
40 70
0
-6
5-
-
-
-
-8 Ensemble percentile (%) the role of an asteroid impact in extinction of non-avian dino
saurs [141], which initially were highly controversial but are now
00
00
00
00
00
00
0
0
0
0
more widely accepted. Ruddiman’s hypothesis, right or wrong, is

0
00
00
00
00
75
50
25
,0
,0
,0
,0
,0
,0
,0
8,
6,
4,
2,
22
20
18
16
14
12
10
Age (yr BP) still controversial. Thus, we minimize this issue by showing aero
sol effects with and without preindustrial human-made aerosols.
Figure 14. Global mean surface temperature change over the past 24 ky,
reproduced from Fig. 2 of Osman et al. [50] including Last Millennium Global aerosols are not monitored with detail needed to define
reanalysis of Tardif et al. [136]. aerosol climate forcing [142, 143]. IPCC12 estimates forcing
.5 .1
0. 0.
−.5 −.1
2
Fe (W/m )
−1.0 −.2
−1.5 −.3
CO2
−2.0 CH4 −.4
N2O −.5
−2.5
Sum of Above Three
−3.0 −.6
20 18 16 14 12 10 8 6 4 2 0 10 8 6 4 2 0
Time (kyBP) Time (kyBP)
Figure 15. GHG climate forcing in past 20 ky with vertical scale expanded for the past 10 ky on the right. GHG amounts are from Schilt et al. [47].
Formulae for forcing are in Supplementary Material.
(Fig. 17a) from assumed precursor emissions, a herculean task due constraints. Unknowns are ECS, net climate forcing (aerosol forc
to many aerosol types and complex cloud effects. Aerosol forcing ing is unmeasured), and ocean mixing (many ocean models are
uncertainty is comparable to its estimated value (Fig. 17a), which too diffusive). Constraints are observed global temperature
is constrained more by observed global temperature change than change and Earth’s energy imbalance (EEI) [80]. Knutti [150] and
by aerosol measurements [144]. IPCC’s best estimate of aerosol Hansen [75] suggest that many climate models compensate for
forcing (Fig. 17) and GHG history define the percent of GHG forcing excessive ocean mixing (which reduces surface warming) by us
offset by aerosol cooling—the dark blue area in Fig. 17b. However, ing aerosol forcing less negative than the real world, thus achiev
if human-made aerosol forcing was −0.5 W/m2 by 1750, offsetting ing realistic surface warming. This issue is unresolved and
þ0.5 W/m2 GHG forcing, this forcing should be included. Such complicated by the finding that cloud feedbacks can buffer ocean
aerosol forcing—largely via effects of land use and biomass fuels heat uptake (Climate response time section), affecting interpreta
on clouds—continues today. Thirty million people in the United tion of EEI.
States use wood for heating [145]. Such fuels are also common in IPCC AR6 WG1 best estimate of aerosol forcing (Table AIII.3)
Europe [146, 147] and much of the world. [12] is near maximum (negative) value by 1975, then nearly con

Figure 17b encapsulates two alternative views of aerosol his stant until rising in the 21st century to –1.09 W/m2 in 2019
tory. IPCC aerosol forcing slowly becomes important relative to (Fig. 18). We use this IPCC aerosol forcing in climate simulations
GHG forcing. In our view, civilization always produced aerosols as here. We also use an alternative aerosol scenario [151] that
well as GHGs. As sea level stabilized, organized societies and popu reaches –1.63 W/m2 in 2010 relative to 1880 and –1.8 W/m2 rela
lation grew as coastal biologic productivity increased [148] and ag tive to 1850 (Fig. 18) based on modeling of Koch [152] that in
riculture developed. Wood was the main fuel. Aerosols travel great cluded changing technology factors defined by Novakov [153].
distances, as shown by Asian aerosols in North America [149]. This alternative scenario11 is comparable to the forcing in some
Humans contributed to both rising GHG and aerosol climate forc current aerosol models (Fig. 18). Human-made aerosol forcing
ings in the past 6000 years. One result is that human-caused aero relative to several millennia ago may be even more negative, by
sol climate forcing is at least 0.5 W/m2 more than usually about –0.5 W/m2 as discussed above, but the additional forcing
assumed. Thus, the Faustian payment that will eventually come was offset by increasing GHGs and thus those additional forcings
due is also larger, as discussed in Summary section. are neglected, with climate assumed to be in approximate equi
librium in 1850.
Ambiguity in aerosol climate forcing Many combinations of climate sensitivity and aerosol forcing
In this section we discuss uncertainty in the aerosol forcing. We can fit observed global warming. The GISS (2014) model (ECS ¼
discuss why global warming in the past century—often used to 2.6� C) with IPCC AR6 aerosol forcing can match observed warm
infer climate sensitivity—is ill-suited for that purpose. ing (Fig. 19) in the last half century (when human-made climate
Recent global warming does not yield a unique ECS because forcing overwhelmed natural forcings, unforced climate variabil
warming depends on three major unknowns with only two basic ity, and flaws in observations). However, agreement also can be
achieved by climate models with high ECS. The GISS (2020) model
(with ECS ¼ 3.5� C) yields greater warming than observed if IPCC
0
IPCC AR6 CMIP6
Forcing (W/m2)
Chapter 7 AR6
−1 GCMs
−2
IPCC Table AIII.3 (1850 = 0)
Alternative Aerosol Scenario
−3 Matrix Aerosol Model (Bauer et al., 2020)
OMA Aerosol Model (Bauer et al., 2020)
1850 1900 1950 2000 2050
Figure 16. Sea level since the last glacial period relative to present. Figure 18. Aerosol forcing relative to 1850 from IPCC AR6, an alternative
Credit: Robert Rohde [138]. aerosol scenario [151] two aerosol model scenarios of Bauer et al. [154].
Figure 17. (a) Estimated greenhouse gas and aerosol forcings relative to 1750 values. (b) Aerosol forcing as percent of GHG forcing. Forcings for dark
blue area are relative to 1750. Light blue area adds 0.5 W/m2 forcing estimated for human-caused aerosols from fires, biofuels and land use.
(a) GISS (2014) Model (b) GISS (2020) Model

Forcing = GHG + IPCC Aerosols
1.0 Forcing = GHG + IPCC Aerosols 1.0 Forcing = GHG + Alternative Aerosol Scenario
Observed Temperature
ECS = 4C (Alternative Aerosol Scenario)
Observed Temperature
Temperature (°C)
.5 .5
0. 0.

−.5 −.5
1850 1900 1950 2000 1850 1900 1950 2000
Figure 19. Global temperature change TG due to aerosols þ GHGs calculated with Green’s function Equation (5) using GISS (2014) and GISS (2020)
response functions (Fig. 4). Observed temperature is the NASA GISS analysis [155, 156]. Base period: 1951–1980 for observations and model.
(a) (b)
Figure 20. Total sulfate (parts per trillion by volume) and percentage of total sulfate provided by shipping in simulations of Jin et al. [157] prior to IMO
regulations on sulfur content of fuels.
aerosol forcing is used, but less than observed for the alternative study aerosol and cloud physics is provided by a recent change in
aerosol scenario (Fig. 19). This latter aerosol scenario achieves the IMO (International Maritime Organization) regulations on
agreement with observed warming if ECS �4� C (green curve in ship emissions.
Fig. 19).12 Agreement can be achieved with even higher ECS by
use of a still more negative aerosol forcing. The great inadvertent aerosol experiment
The issue we raise is the magnitude of the aerosol forcing, Sulfate aerosols are cloud condensation nuclei (CCN), so sulfate
with implications for future warming when particulate air pollu emissions by ships result in a larger number of smaller cloud par
tion is likely to be reduced. We suggest that IPCC reports may ticles, thus affecting cloud albedo and cloud lifetime [144]. Ships
have gravitated toward climate sensitivity near 3� C for 2 � CO2 in provide a large percentage of sulfates in the North Pacific and
part because of difficulty that models have in realistically simu North Atlantic regions (Fig. 20). It has been suggested that cooling
lating amplifying cloud feedbacks and a climate model tendency by these clouds is overestimated because of cloud liquid water
for excessive mixing of heat into the deep ocean. Our finding adjustments [159], but Manshausen et al. [160] present evidence
from paleoclimate analysis that ECS is 1.2� C ± 0.3� C per W/m2 that liquid water path (LWP) effects are substantial even in
(4.8� C ± 1.2� C for 2 � CO2) implies that the (unmeasured) aerosol regions without visible ship-tracks; they estimate a LWP forcing
forcing must be more negative than IPCC’s best estimate. In −0.76 ± 0.27 W/m2, in stark contrast with the IPCC estimate of
turn—because aerosol-cloud interactions are the main source of þ0.2 ± 0.2 W/m2. Wall et al. [161] use satellite observations to
uncertainty in aerosol forcing—this finding emphasizes the need quantify relationships between sulfates and low-level clouds;
to measure both global aerosol and cloud particle properties. they estimate a sulfate indirect aerosol forcing of −1.11 ± 0.43 W/
The case for monitoring global aerosol climate forcing will m2 over the global ocean. The range of aerosol forcings used in
grow as recognition of the need to slow and reverse climate CMIP6 and AR6 GCMs (small blue bar in Fig. 18) is not a measure
change emerges. Aerosol and cloud particle microphysics must of aerosol forcing uncertainty. The larger bar, from Chapter 7
be measured with precision adequate to define the forcing [142, [162] of AR6, has negative forcing as great as –2 W/m2, but even
158]. In the absence of such Keeling-like global monitoring, that does not measure the full uncertainty.
progress can be made via more limited satellite measurements of Changes of IMO emission regulations provide a great opportu
aerosol and cloud properties, field studies, and aerosol and nity for insight into aerosol climate forcing. Sulfur content of
cloud modeling. As described next, a great opportunity to fuels was limited to 1% in 2010 near the coasts of North America
and in the North Sea, Baltic Sea and English Channel, and further contributions to the 20-year trend of absorbed solar energy show
restricted there to 0.1% in 2015 [163]. In 2020 a limit of 0.5% was that clouds provide most of the change. Surface albedo decrease
imposed worldwide. The 1% limit did not have a noticeable effect due to sea ice decline contributes to the 20-year trend in the
on ship-tracks, but a striking reduction of ship-tracks was found Northern Hemisphere, but that sea ice decline occurred espe
after the 2015 IMO regulations, especially in the regions near cially in 2007, with minimum sea ice cover reached in 2012; over
land where emissions were specifically limited [164]. Following the past decade as global and hemispheric albedos declined, sea
the additional 2020 regulations [165], global ship-tracks were re ice had little trend [169]. Potential causes of the cloud changes in
duced more than 50% [166]. clude: (1) reduced aerosol forcing, (2) cloud feedbacks to global
Earth’s albedo (reflectivity) measured by CERES (Clouds and warming, (3) natural variability [170]. Absorbed solar energy was
Earth’s Radiant Energy System) satellite-borne instruments [81] 0.77 W/m2 greater in Jan2015-Dec2022 than in the first decade of
over the 22-years March 2000 to March 2022 reveal a decrease of CERES data at latitudes 20–60� S (Fig. 22), a region of relatively lit
albedo and thus an increase of absorbed solar energy coinciding tle ship traffic. This change is an order of magnitude larger than
with the 2015 change of IMO emission regulations. Global the estimate of potential detector degradation [81].

absorbed solar energy is þ 1.05 W/m2 in the period January 2015 Climate models predict a reduction of cloud albedo in this re
through December 2022 relative to the mean for the first 10 years gion as a feedback effect driven by global warming [12] (Sec.
of data (Fig. 21). This increase is 5 times greater than the stan 7.4.2.4). Continued monitoring of absorbed energy can confirm
dard deviation (0.21 W/m2) of annual absorbed solar energy in the reality of the change, but without global monitoring of de
the first 10 years of data and 4.5 times greater than the standard tailed physical properties of aerosols and clouds [142], it will be
deviation (0.23 W/m2) of CERES data through December 2014. difficult to apportion observed change among candidate causes.
The increase of absorbed solar energy is notably larger than esti North Pacific and North Atlantic regions of heavy ship traffic
mated potential CERES instrument drift, which is <0.085 W/m2 are ripe for detailed study of cloud changes and their causes, al
per decade [81]. Increased solar energy absorption occurred de
though unforced cloud variability is large in such sub-global
spite 2015–2020 being the declining phase of the �11-year solar
regions. Both regions have increased absorption of solar radiation
irradiance cycle [167]. Nor can increased absorption be attributed
after 2015 (Fig. 22). The 2014–2017 maximum absorption in the
to correlation of Earth’s albedo (and absorbed solar energy) with
North Pacific is likely enhanced by reduced cloud cover during
the Pacific Decadal Oscillation (PDO): the PDO did shift to the
the positive PDO, but the more recent high absorption is during
positive phase in 2014–2017, but it returned to the negative phase
the negative PDO phase. In the North Atlantic, persistence of in
in 2017–2022 [168].
creased absorption for several years exceeds prior variability, but
Given the large increase of absorbed solar energy, cloud
longer records plus aerosol and cloud microphysical data are
changes are likely the main cause. Quantitative analysis [168] of
needed for interpretation.
Global Absorbed Solar Radiation Anomaly (W/m2)

3 Summary
Monthly mean
12−month mean Climate change is characterized by delayed response and ampli
2 Jan 2015 − Dec 2019 mean
Jan 2020 − Jun 2023 mean fying feedbacks. Delayed response makes human-made climate
1.29
1 0.99 forcing a threat to today’s public and future generations because
of the practical difficulty of reversing the forcing once conse
0
quences become apparent. Feedbacks determine climate sensi
−1
tivity to any applied forcing. We find that Earth’s climate is very
sensitive—more sensitive than the best estimate of the
2000 2005 2010 2015 2020 2025
Intergovernmental Panel on Climate Change (IPCC)—which
Figure 21. Global absorbed solar radiation (W/m2) relative to mean of implies that there is a great amount of climate change ‘in the
the first 120 months of CERES data. CERES data are available at http:// pipeline.’ Extraordinary actions are needed to reduce the net
ceres.larc.nasa.gov/order_data.php. human-made climate forcing, as is required to reduce global
(a) Absorbed Solar Radiation Anomaly (b) Absorbed Solar Radiation Anomaly
2.5 5
Global North Pacific Ocean
2.0 Southern Hemisphere 20-60°S (largely ocean) 4
North Atlantic Ocean 3.60
1.5 3
Anomaly (W/m2)
1.29 2.64
2.10
1.0 0.99 2
0.89
0.67
.5 1
0.74
0. 0
−.5 −1
−1.0 −2
−1.5 −3
2000 2005 2010 2015 2020 2025 2000 2005 2010 2015 2020 2025
Figure 22. Absorbed solar radiation for indicated regions relative to first 120 months of CERES data. Southern Hemisphere 20–60� S is 89% ocean. North
Atlantic is (20–60� N, 0–60� W) and North Pacific is (20–60� N, 120–220� W). Data source: http://ceres.larc.nasa.gov/order_data.php.
warming and avoid highly undesirable consequences for human remained about 100 years. There are two reasons for this: one
ity and nature. that is obvious and one that is more interesting and informative.
The surface in the newer model warms as fast as in the older
Equilibrium climate sensitivity (ECS) model, but it must achieve greater warming to reach 63% of equi
The 1979 Charney study [4] considered an idealized climate sen librium because its ECS is higher, which is one reason that the re
sitivity in which ice sheets and non-CO2 GHGs are fixed. The sponse time remains long. The other reason is that Earth’s
Charney group estimated that the equilibrium response to 2 � energy imbalance (EEI) in the newer model decreases rapidly. EEI
CO2, a forcing of 4 W/m2, was 3� C, thus an ECS of 0.75� C per W/ defines the rate that heat is pumped into the ocean, so a smaller
m2, with one standard deviation uncertainty r ¼ 0.375� C. EEI implies a longer time for the ocean to reach its new equilib
Charney’s estimate stood as the canonical ECS for more than 40 rium temperature. Quick drop of EEI—in the first year after intro
years. The current IPCC report [12] concludes that 3� C for 2 � duction of the forcing—implies existence of ultrafast feedback in
CO2 is their best estimate for ECS. the GISS (2020) model. For want of an alternative with such a
We compare recent glacial and interglacial climates to infer large effect on Earth’s energy budget, we infer a rapid cloud feed

ECS with a precision not possible with climate models alone. back and we suggest (Slow, fast and ultrafast feedbacks section)
Uncertainty about Last Glacial Maximum (LGM) temperature has a set of brief GCM runs that define cloud changes and other diag
been resolved independently with consistent results by Tierney nostic quantities to an arbitrary accuracy.
et al. [49] and Seltzer et al. [51]. The Tierney approach, using a col The Charney report [4] recognized that clouds were a main
lection of geochemical temperature indicators in a global analy cause of a wide range in ECS estimates. Today, clouds still cast
sis constrained by climate change patterns defined by a global uncertainty on climate predictions. Several CMIP6 [34] GCMs
climate model, is used by Osman et al. [50] to find peak LGM cool have ECS of �4–6� C for 2�CO2 [171, 172] with the high sensitivity
ing 7.0 ± 1� C (2r, 95% confidence) at 21–18 kyBP. We show that, caused by cloud feedbacks [84]. As cloud modeling progresses, it
accounting for polar amplification, these analyses are consistent will aid understanding if climate models report their 2 � CO2 re
with the 5.8 ± 0.6� C LGM cooling of land areas between 45� S and sponse functions for both temperature and EEI (Earth’s en
35� N found by Seltzer et al. using the temperature-dependent sol ergy imbalance).
ubility of dissolved noble gases in ancient groundwater. The forc Fast EEI response—faster than global temperature response—
has a practical effect: observed EEI understates the reduction of
ing that maintained the 7� C LGM cooling was the sum of 2.25 ±
climate forcing required to stabilize climate. Although the mag
0.45 W/m2 (2r) from GHGs and 3.5 ± 1.0 W/m2 (2r) from the LGM
nitude of this effect is uncertain (see Supplementary Material
surface albedo, thus 5.75 ± 1.1 W/m2 (2r). ECS implied by the
SM6), it makes the task of restoring a hospitable climate and sav
LGM is thus 1.22 ± 0.29� C (2r) per W/m2, which, at this final step,
ing coastal cities more challenging. On the other hand, long cli
we round to 1.2 ± 0.3� C per W/m2. For transparency, we have
mate response time implies the potential for educated policies to
combined uncertainties via simple RMS (root-mean-square). ECS
affect the climate outcome before the most undesirable conse
as low as 3� C for 2 � CO2 is excluded at the 3r level, i.e. with
quences occur.
99.7% confidence.
The time required for climate to reach a new equilibrium is
More sophisticated mathematical analysis, which has merits
relevant to policy (Perspective on policy implications section),
but introduces opportunity for prior bias and obfuscation, is not
but there is another response time of practical importance. With
essential; error assessment ultimately involves expert judgment.
climate in a state of disequilibrium, how much time do we have
Instead, focus is needed on the largest source of error: LGM sur
before we pass the point of no return, the point where major cli
face albedo change, which is uncertain because of the effect of
mate impacts are locked in, beyond our ability to control? That’s
cloud shielding on the efficacy of the forcing. As cloud modeling
a complex matter; it requires understanding of ‘slow’ feedbacks,
is advancing rapidly, this topic is ripe for collaboration of CMIP
especially ice sheets. It also depends on how far climate is out of
[53] (Coupled Model Intercomparison Project) with PMIP [54] equilibrium. Thus, we first consider the full Earth system
(Paleoclimate Modelling Intercomparison Project). Simulations sensitivity.
should include at the same time change of surface albedo and to
pography of ice sheets, vegetation change, and exposure of conti Earth system sensitivity (ESS)
nental shelves due to lower sea level. The Cenozoic era—the past 66 million years—provides an oppor
Knowledge of climate sensitivity can be advanced further via tunity to study Earth system sensitivity via a consistent analysis
analysis of the wide climate range in the Cenozoic era (Earth sys for climate ranging from hothouse conditions with Earth 15� C
tem sensitivity section). However, interpretation of data and warmer and sea level 60 m higher than preindustrial climate to
models, and especially projections of climate change, depend on glacial conditions with Earth 7� C cooler and sea level 120 m lower
understanding of climate response time. than preindustrial. Atmospheric CO2 amount in the past 800 000
years (Fig. 2), confirms expectation that CO2 is the main control
Climate response time knob [87] on global temperature. We can assume this control
We expected climate response time—the time for climate to ap existed when CO2 amount varied due to CO2 emissions caused by
proach a new equilibrium after imposition of a forcing—to be plate tectonics (continental drift). The two-step [91] that the
come faster as mixing of heat in ocean models improved [75]. Indian plate executed as it moved through the Tethys (now
That expectation was not met when we compared two genera Indian) ocean left a signature in atmospheric CO2 and global tem
tions of the GISS GCM (global climate model). The GISS (2020) perature. CO2 emissions from subduction of ocean crust were
GCM is improved [32, 33] in its ocean simulation over the GISS greatest when the Indian plate was moving fastest (inset, Fig. 6)
(2014) GCM as a result of higher vertical and horizontal resolu and peaked at its hard collision with the Eurasian plate at 50
tion, more realistic parameterization of sub-grid scale motions, MyBP. Diminishing metamorphic CO2 emissions continue as the
and correction of errors in the ocean computer program [32]. Yet Indian plate is subducted beneath the Eurasian plate, pushing up
the time for the model to achieve 63% of its equilibrium response the Himalayan Mountains, but carbon drawdown from
weathering and burial of organic carbon exceeds emissions. persuasive explanation for absence of global warming in the last
Motion of the Indian Plate thus dominates the broad sweep of half of the Holocene (Fig. 14) as GHG forcing increased 0.5 W/m2
Cenozoic CO2, but igneous provinces play a role. The North (Fig. 15). Climate models without a growing negative aerosol forc
Atlantic Igneous Province (caused by a rift in the sea floor as ing yield notable warming in that period [173], a warming that, in
Greenland pulled away from Europe) that triggered the fact, did not occur. Negative aerosol forcing, increasing as civili
Paleocene-Eocene Thermal Maximum (PETM) event about 56 zation developed and population grew, is expected. As humans
MyBP and the Columbia River Flood Basalt about 15 MyBP (Fig. 6) burned fuels at a growing rate—wood and other biomass for mil
are most notable. lennia and fossil fuels in the industrial era—aerosols as well as
We infer the Cenozoic history of sea surface temperature GHGs were an abundant, growing, biproduct. The aerosol source
(SST) at sites of deepwater formation from the oxygen isotope from wood-burning has continued in modern times [146]. GHGs
d18O in shells of deep-ocean-dwelling foraminifera preserved in are long-lived and accumulate, so their forcing dominates even
ocean sediments [44, 90]. High latitude SST change—including a tually, unless aerosol emissions grow higher and higher—the
correction term as SST approaches the freezing point—provides Faustian bargain [98].

an accurate estimate of global surface temperature change. This Multiple lines of evidence show that aerosol forcing peaked
Cenozoic temperature history and climate sensitivity inferred early this century [174]. Emissions from the largest sources,
from the LGM cooling yield an estimate of Cenozoic CO2 history. China and India, were increasing in 2000, but by 2010 when the
We suggest that this whole-Cenozoic approach may define the first limits on ship emissions were imposed, China’s emissions
CO2 history (Fig. 9a) more accurately than CO2 proxy measure were declining. We estimate peak (negative) aerosol forcing as at
ments. We find CO2 about 325 ppm in the early Pliocene and 450 least 1.5–2 W/m2, with turning point at 2010, consistent with
ppm at transition to glaciated Antarctica. Global climate models Fig. 3 of Bauer et al. [175] GHG plus aerosol forcing grew þ0.3 W/
(GCMs) that isolate on the Pliocene tend to use CO2 levels of order m2 per decade (GHGs: þ0.45, aerosols: –0.15) during 1970–2010,
400 ppm in attempts to match actual Pliocene warmth and ice which produced warming of 0.18� C per decade. With current pol
sheet models use CO2 of order 700 ppm or greater to achieve ice icies, we expect climate forcing for a few decades post-2010 to in
sheet disintegration on Antarctica, which suggests that the mod crease 0.5–06 W/m2 per decade and produce global warming of at
els are not realistically capturing amplifying feedback processes least þ0.27� C per decade. In that case, global warming will reach
(see Cenozoic CO2 section). 1.5� C in the 2020s and 2� C before 2050 (Fig. 24). Such acceleration
The Cenozoic provides a perspective on present greenhouse is dangerous in a climate system that is already far out of equilib
gas (GHG) levels. The dashed line in Fig. 23 is the ‘we are here’ rium and dominated by multiple amplifying feedbacks.
level of GHG climate forcing. Today’s GHG forcing of 4.6 W/m2 is The sharp change of ship emissions in 2020 (The great inad
relative to mid-Holocene CO2 of 260 ppm; we present evidence in vertent aerosol experiment section) provides an indirect measure
Cenozoic CO2 section that 260 ppm is the natural Holocene CO2 of aerosol effects. Diamond [176] finds a cloud brightness de
level. Human-caused GHG forcing today is already above the crease of order 1 W/m2 in a shipping corridor. We find a larger ef
level needed to deglaciate Antarctica, if such forcing is left in fect, increased absorption of about 3 W/m2 in regions of heavy
place long enough. We do not predict full deglaciation of ship traffic in the North Atlantic and North Pacific (Fig. 22), but a
Antarctica on a time scale people care about—rather we draw at longer record is needed to define significance. However, the sin
tention to how far today’s climate is out of equilibrium with gle best sentinel for global climate change is Earth’s en
today’s GHG level. This is one measure of how strongly humanity ergy imbalance.
is pushing the climate system. Stabilizing climate requires re
moving the disequilibrium by reducing human-made climate Earth’s energy imbalance
forcing. A danger is that it will become difficult or implausible to Earth’s energy imbalance (EEI) is the net gain (or loss) of energy
prevent large sea level rise, if deglaciation is allowed to get by the planet, the difference between absorbed solar energy and
well underway. emitted thermal (heat) radiation. As long as EEI is positive, Earth
GHGs are not the only large human-made climate forcing. will continue to get hotter. EEI is hard to measure, a small differ
Understanding of ongoing climate change requires that we also ence between two large quantities (Earth absorbs and emits
include the effect of aerosols (fine airborne particles).
Aerosols 2.4
2.2 12−month Running Mean
Aerosol climate forcing is larger than the IPCC AR6 estimate and Accelerated
2.0 132−month Running Mean
has likely been significant for millennia. We know of no other January−December Mean Warming
1.8
Temperature Anomaly (°C)
Best Linear Fit (1970−2010)

↓
1.6 (0.18°C/decade)
1.4
Super →
1.2 El Ninos
1.0 ↓
.8
.6
.4
.2
0.
−.2
1880 1900 1920 1940 1960 1980 2000 2020 2040
Figure 24. Global temperature relative to 1880–1920. Edges of the

Figure 23. Forcing required to yield Cenozoic temperature for today’s predicted post-2010 accelerated warming rate (see text) are 0.36 and
solar irradiance, compared with human-made GHG forcing in 2022. 0.27� C per decade.
about 240 W/m2 averaged over the entire planetary surface), but Equilibrium warming versus committed warming
change of EEI can be well-measured from space [81]. Absolute Equilibrium warming for today’s climate forcing is the warming re
calibration is from the change of heat in the heat reservoirs, quired to restore Earth’s energy balance if atmospheric composi
mainly the global ocean, over a period of at least a decade, as tion is fixed at today’s conditions. Equilibrium warming is a
needed to reduce error due to the finite number of places that benchmark that can be evaluated from atmospheric composition
the ocean is sampled [80]. EEI varies year-to-year (Fig. 25), largely and paleoclimate data, with little involvement of climate models.
because global cloud amount varies with weather and ocean dy It is the standard benchmark used in definition of the Charney ECS
namics, but averaged over several years EEI helps inform us (equilibrium climate sensitivity excluding slow feedbacks) [4] and
about what is needed to stabilize climate. ESS (Earth system sensitivity, which includes slow feedbacks such
The data indicate that EEI has doubled since the first decade as ice sheet size) [71]. GHG climate forcing now is 4.6 W/m2 relative
of this century (Fig. 25). This increase is one basis for our predic to the mid-Holocene (7 kyBP) or 4.1 W/m2 relative to 1750. There is
tion of post-2010 acceleration of the global warming rate. The EEI little merit in debating whether GHG forcing is 4.6 or 4.1 W/m2 be
increase may be partly due to restrictions on maritime aerosol cause it is still increasing 0.5 W/m2 per decade (Perspective on pol

precursor emissions imposed in 2015 and 2020 (The great inad icy implications section). ECS response to 4.6 W/m2 forcing for
vertent aerosol experiment section), but the growth rate of GHG climate sensitivity 1.2� C per W/m2 is 5.5� C. The eventual Earth sys
climate forcing also increased in 2015 and since has remained at tem response (ESS) to sustained 4.6 W/m2 forcing is about 10� C
the higher level (Equilibrium warming versus committed warm (Earth system sensitivity section), because that forcing is large
ing section). enough to deglaciate Antarctica (Fig. 23). Net human-made forcing
Reduction of climate forcing needed to reduce EEI to zero is today is probably near 3 W/m2 due to negative aerosol forcing.
greater than EEI because of ultrafast cloud feedback (Slow, fast Even 3 W/m2 may be sufficient to largely deglaciate Antarctica, if
and ultrafast feedbacks section), but the magnitude of this effect the forcing is left in place permanently (Fig. 23).
is uncertain (SM6). Cloud feedbacks are only beginning to be sim ‘Committed warming’ is less precisely defined; even in the
ulated well, but climate sensitivity near 1.2� C per W/m2 implies current IPCC report [12] (p. 2222) it has multiple definitions. One
that the net cloud feedback is large and deserves greater atten concept is the warming that occurs if human-made GHG emis
tion. Precise monitoring of EEI is essential as a sentinel for future sions cease today, but that definition is ill-posed as well as unre
climate change and to assess efforts to stabilize climate and alistic. Do aerosol emissions also cease? That would cause a
avoid undesirable consequences. Global satellite monitoring of sudden leap in Earth’s energy imbalance, a ‘termination shock,’
geographical and temporal changes of EEI and ocean in situ moni as the cooling effect of human-made aerosols disappears. A more
toring (especially in polar regions of rapid change) are both useful definition is the warming that will occur with plausibly
needed for the sake of understanding ongoing climate change. rapid phasedown of GHG emissions, including comparison with
ongoing reality. However, the required ‘integrated assessment
models,’ while useful, are complex and contain questionable
Earth’s Energy Imbalance (W/m2) assumptions that can mislead policy (see Perspective on policy
3
Monthly implications section).
12−Month Running Mean
Jan 2020 − June 2023 Nature’s capacity for restoration provides hope that future
2 Jan 2015 − Dec 2019 warming can be limited, if humanity moves promptly toward
1.36 sustainable energy and climate policies. Earth’s ability to remove
1 1.12 human-made CO2 emissions from the atmosphere is revealed by
Fig. 26. Fossil fuel emissions now total more than 10 GtC/year,
0 which is almost 5 ppm of CO2, yet CO2 in the air is only increasing
2.5 ppm/year. The other half is being taken up by the ocean, solid
2000 2005 2010 2015 2020 2025 land, and biosphere. Indeed, Earth is taking up even more be
cause deforestation, fires, and poor agricultural and forestry
Figure 25. 12-month running-mean of Earth’s energy imbalance
from CERES satellite data [81] normalized to 0.71 W/m2 mean for practices are additional human-made CO2 sources. If human
July 2005–June 2015 (blue bar) from in situ data [80]. emissions ceased, atmospheric CO2 would initially decline a few
(a) (b)
Figure 26. Fossil fuel emissions divided into portions appearing in the annual increase of airborne CO2 and the remainder, which is taken up by the
ocean and land (1 ppm CO2 � 2.12 GtC).
ppm per year, but uptake would soon slow—it would take millen Fossil fuels still provide most of the world’s energy (Fig. 27a) and
nia for CO2 to reach preindustrial levels [131]. This underscores produce most CO2 emissions (Fig. 27b). Much of the world is still
the difficulty of restoring Earth’s energy balance via emission in early or middle stages of economic development. Energy is
reductions alone. Furthermore, fossil fuels have raised living needed and fossil fuels are a convenient, affordable source of en
standards in most of the world and still provide 80% of the ergy. One gallon (3.8 l) of gasoline (petrol) provides the work
world’s energy, which contributes to a policy inertia. As the real equivalent of more than 400 h labor by a healthy adult. These
ity of climate change emerges, the delayed response of climate benefits are the basic reason for continued high emissions. The
and amplifying feedbacks assure that the world has already set Covid pandemic dented emissions in 2020, but 2022 global emis
sail onto even more turbulent climate seas. Scientists must do sions were a record high level. Fossil fuel emissions from mature
their best to help the public understand policy options that may economies are beginning to fall due to increasing energy effi
preserve and restore a propitious climate for future generations. ciency, introduction of carbon-free energies, and export of
manufacturing from mature economies to emerging economies.
However, at least so far, those reductions have been more than
Perspective on policy implications

offset by increasing emissions in developing nations (Fig. 28).
This section is the first author’s perspective based on more than The potential for rising CO2 to be a serious threat to humanity
20 years of experience on policy issues that began with a paper was the reason for the 1979 Charney report, which confirmed
[179] and two workshops [180] that he organized at the East-West that climate was likely sensitive to expected CO2 levels in the
Center in Hawaii, followed by meetings and workshops with util 21st century. In the 1980s it emerged that high climate sensitivity
ity experts and trips to more than a dozen nations for discussions implied a long delay between changing atmospheric composition
with government officials, energy experts, and environmental and the full climate response. Ice core data revealed the impor
ists. The aim was to find a realistic scenario with a bright energy tance of amplifying climate feedbacks. A climate characterized
and climate future, with emphasis on cooperation between the by delayed response and amplifying feedbacks is especially dan
West and nations with emerging or underdeveloped economies. gerous because the public and policymakers are unlikely to make
fundamental changes in world energy systems until they see visi
Energy, CO2 and the climate threat ble evidence of the threat. Thus, it is incumbent on scientists to
The world’s energy and climate path has good reason: fossil fuels make this situation clear to the public as soon as possible. That
powered the industrial revolution and raised living standards. task is complicated by the phenomenon of scientific reticence.
(a) Global Energy Consumption (b) Global CO 2 Emissions

600 10
Energy (Gt oil equivalent / year)
14
Renewables Gas Flaring
Hydroelectric Cement Production
Energy (Exajoule/year)
12 500
Emissions (GtC/year)
Nuclear 8 Gas
10 Gas 400 Oil
Oil 6 Coal
8
Coal 300
6 4
200
4
100 2
2
0 0 0
1900 1920 1940 1960 1980 2000 2020 1900 1920 1940 1960 1980 2000 2020
Figure 27. Global energy consumption and CO2 emissions (Hefner at al. [177] and Energy Institute [178]).
Global Emissions (GtC/Year)

10.0
Global 10
Mature Economies
Emerging Economies 8
1.0 6
Emerging
Economies
4
2
.1 Mature Economies
0
1850 1900 1950 2000 1900 1920 1940 1960 1980 2000 2020
Figure 28. Fossil fuel CO2 emissions from mature and emerging economies. China is counted as an emerging economy. Data sources as in Fig. 27.
Scientific reticence century. The 15 authors, representing leading GCM groups, used
Bernard Barber decried the absence of attention to scientific reti 21 climate projections from eight ‘ … state-of-the-science, IPCC
cence, a tendency of scientists to resist scientific discovery or class … ’ GCMs to conclude that ‘ … the probability of an AMOC
new ideas [139]. Richard Feynman needled fellow physicists collapse is negligible. This is contrary to a recent modeling study
about their reticence to challenge authority [181], specifically to [Hansen et al., 2016] that used a much larger, and in our assess
correct the electron charge that Millikan derived in his famous ment unrealistic, Northern Hemisphere freshwater forcing …
oil drop experiment. Later researchers moved Millikan’s result According to our probabilistic assessment, the likelihood of an
bit by bit—experimental uncertainties allow judgment—reaching AMOC collapse remains very small (<1% probability) if global
an accurate result only after years. Their reticence embarrassed warming is below � 5K … ’[189]. They treated the ensemble of
the physics community but caused no harm to society. A factor their model results as if it were the probability distribution for
that may contribute to reticence among climate scientists is the real world.
‘delay discounting:’ preference for immediate over delayed In contrast, we used paleoclimate evidence, global modeling,
rewards [182]. The penalty for ‘crying wolf’ is immediate, while and ongoing climate observations. Paleoclimate data [190]

the danger of being blamed for ‘fiddling while Rome was burning’ showed that AMOC shutdown is not unusual and occurred in the
is distant. One of us has noted [183] that larding of papers and Eemian (when global temperature was similar to today), and also
proposals with caveats and uncertainties increases chances of that sea level in the Eemian rose a few meters within a century
obtaining research support. ‘Gradualism’ that results from reti [191] with the likely source being collapse of the West Antarctic
cence is comfortable and well-suited for maintaining long-term ice sheet. Although we would not assert that our model corrected
support. Gradualism is apparent in IPCC’s history in evaluating all excessive ocean mixing, the higher vertical resolution and im
climate sensitivity as summarized in our present paper. Barber proved mixing increased the sensitivity to freshwater flux, as
identifies professional specialization—which causes ‘outsiders’ confirmed in later tests [192]. Modern observations showed and
to be ignored by ‘insiders’—as one cause of reticence; specializa continue to add evidence that the overturning Southern Ocean
tion is relevant to ocean and ice sheet dynamics, matters upon [193, 194] and North Atlantic [195] are already slowing. Growth of
which the future of young people hangs. meltwater injection onto the Southern [196] and North Atlantic
Discussion [184] with field glaciologists13 20 years ago revealed Oceans [197] is consistent with a doubling time of 10–20 years.
frustration with IPCC’s ice sheet assessment. One glaciologist High climate sensitivity inferred in our present paper also implies
said—about a photo [185] of a moulin (a vertical shaft that carries there will be a greater increase of precipitation on polar oceans
meltwater to the base of the Greenland ice sheet)—‘the whole ice than that in most climate models.
sheet is going down that damned hole!’ Concern was based on The indictment of Ice Melt by Bakker et al. [189] was accepted
observed ice sheet changes and paleoclimate evidence of sea by the research community. Papers on the same topics ignored
level rise by several meters in a century, implying that ice sheet our paper or referred to it parenthetically with a note that we
collapse is an exponential process. Thus, as an alternative to ice used unrealistic melt rates, even though these were based on
sheet models, we carried out a study described in Ice Melt [13]. In observations. Ice Melt was blackballed in IPCC’s AR6 report, which
a GCM simulation, we added a growing freshwater flux to the is a form of censorship [14]. Science usually acknowledges alter
ocean surface mixed layer around Greenland and Antarctica, native views and grants ultimate authority to nature. In the opin
with the flux in the early 21st century based on estimates from in ion of our first author, IPCC does not want its authority
situ glaciological studies [186] and satellite data on sea level challenged and is comfortable with gradualism. Caution has
trends near Antarctica [187]. Doubling times of 10 and 20 years merits, but the delayed response and amplifying feedbacks of cli
were used for the growth of freshwater flux. One merit of our mate make excessive reticence a danger. Our present paper—via
GCM was reduced, more realistic, small-scale ocean mixing, with revelation that the equilibrium response to current atmospheric
a result that Antarctic Bottom Water formed close to the composition is a nearly ice-free Antarctica—amplifies concern
Antarctic coast [13], as in the real world. Growth of meltwater about locking in nonlinearly growing sea level rise. Also, our con
and GHG emissions led to shutdown of the North Atlantic and clusion that CO2 was about 450 ppm at Antarctic glaciation dis
Southern Ocean overturning circulations, amplified warming at parages ice sheet models. Portions of the ice sheets may be
the foot of the ice shelves that buttress the ice sheets, and other recalcitrant to rapid change, but enough ice is in contact with the
feedbacks consistent with ‘nonlinearly growing sea level rise, ocean to provide of the order of 25 m (80 feet) of sea level rise.
reaching several meters in 50–150 years’ [13]. Shutdown of ocean Thus, if we allow a few meters of sea level rise, we may lock in
overturning circulation occurs this century, as early as midcen much larger sea level rise.
tury. The 50–150-year time scale for multimeter sea level rise is
consistent with the 10–20-year range for ice melt doubling time. Climate change responsibilities
Real-world ice melt will not follow a smooth curve, but its growth The industrial revolution began in the U.K., which was the largest
rate is likely to accelerate in coming years due to increasing heat source of fossil fuel emissions in the 19th century (Fig. 29a), but
flux into the ocean (Fig. 25). development soon moved to Germany, the rest of Europe, and
We submitted Ice Melt to a journal that makes reviews publicly the U.S. Nearly half of global emissions were from the U.S. in the
available [188]. One reviewer, an IPCC lead author, seemed intent early 20th century, and the U.S. is presently the largest source of
on blocking publication, while the other reviewer described the cumulative emissions (Fig. 29b) that drive climate change [198,
paper as a ‘masterwork of scholarly synthesis, modeling virtuos 199]. Mature economies, mainly in the West, are responsible for
ity, and insight, with profound implications’. Thus, the editor most cumulative emissions, especially on a per capita basis
obtained additional reviewers, who recommended publication. (Fig. 30). Growth of emissions is now occurring in emerging econ
Promptly, an indictment was published [189] of our conclusion omies (Figs 28 and 29a). China’s cumulative emissions will even
that continued high GHG emissions would cause shutdown of tually pass those of the U.S. in the absence of a successful effort
the AMOC (Atlantic Meridional Overturning Circulation) this to replace coal with carbon-free energy.
(a) Emissions by Region

100 Ships/Air
Africa
S&Cent Am
80 Middle East
Asia Pacific
Fraction (%)
Rus India
sia
60 U.S.A. Jap China
Ca an
n&
Au
s
40
United Kingdom Germany
20 Rest of Europe
0
1800 1820 1840 1860 1880 1900 1920 1940 1960 1980 2000 2020
(b) Cumulative Emissions by Region Ships/Air

100 Africa
S&Cent Am
Middle East
Asia Pacific
80 Rus Chin
a India
Fraction (%)
sia
U.S.A.
60 Japan
Can&Aus
40
United Kingdom Germany Rest of Europe
20
0
1800 1820 1840 1860 1880 1900 1920 1940 1960 1980 2000 2020
Figure 29. Fossil fuel CO2 emissions by nation or region as a fraction of global emissions. Data sources as in Fig. 27.
.08
O3
MPTGs+OTGs RCP8.5
.06 N 2O
DFe (W/m2/year)
CH4
CO2
.04 Gap ~ 0.030 W/m2
.02 RCP6.0 RCP4.5
RCP2.6
.00
1960 1980 2000 2020 2040
Figure 30. Cumulative per capita national fossil fuel emissions [200].
Figure 31. Annual growth of climate forcing by GHGs [38] including part
of O3 forcing not included in CH4 forcing (Supplementary Material).
Greenhouse gas emissions situation MPTG and OTG are Montreal Protocol and Other Trace Gases.
The United Nations uses a target for maximum global warming
to cajole progress in limiting climate change. The 2015 Paris reduce emissions as rapidly as practical and shows that carbon
Agreement [201] aimed to hold ‘the increase in the global average capture cannot be viewed as the solution, although it may play a
temperature to well below 2� C above the pre-industrial levels role in a portfolio of policies, if its cost is driven down.
and pursue efforts to limit the temperature increase to 1.5� C IPCC (Intergovernmental Panel on Climate Change), the sci
above the pre-industrial levels.’ The IPCC AR5 report added a cli entific body advising the world on climate, has not bluntly in
mate forcing scenario, RCP2.6, with a rapid decrease of GHG cli formed the world that the present precatory policy approach
mate forcings, as needed to prevent global warming from will not keep warming below 1.5� C or even 2� C. The ‘tragedy of
exceeding 2� C. Since then, a gap between that scenario and real the commons’ [204] is that, as long as fossil fuel pollution can
ity opened and is growing (Fig. 31). The 0.03 W/m2 gap in 2022 be dumped in the air free of charge, agreements such as the
could be closed by extracting CO2 from the air. However, required Kyoto Protocol [205] and Paris Agreement have limited effect on
negative emissions (CO2 extracted from the air and stored per global emissions. Political leaders profess ambitions for dubious
manently) must be larger than the desired atmospheric CO2 re net-zero emissions while fossil fuel extraction expands. IPCC
duction by a factor of about 1.7 [63]. Thus, the required CO2 scenarios that phase down human-made climate change
extraction is 2.1 ppm, which is 7.6 GtC. Based on a pilot direct-air amount to ‘a miracle will occur’. The IPCC scenario that moves
carbon capture plant, Keith [202] estimates an extraction cost of rapidly to negative global emissions (RCP2.6) has vast biomass-
$450–920 per tC, as clarified elsewhere [203]. Keith’s cost range burning powerplants that capture and sequester CO2, a nature-
yields an extraction cost of $3.4–7.0 trillion. That covers excess ravaging, food-security-threatening [206], proposition without
emissions in 2022 only; it is an annual cost. Given the difficulty scientific and engineering credibility and without a realistic
the UN faced in raising $0.1 trillion for climate purposes and the chance of being deployed at scale and on time to address the cli
growing emissions gap (Fig. 31), this example shows the need to mate threat.
Climate and energy policy Thus, a rising price on GHG emissions is needed, enforced by bor
Climate science reveals the threat of being too late. ‘Being too der duties on products from nations without a carbon fee. Public
late’ refers not only to warning of the climate threat, but also to buy-in and maximum efficacy require the funds to be distributed
technical advice on policy implications. Are we scientists not to the public, which will also address wealth disparity.
complicit if we allow reticence and comfort to obfuscate our de Economists in the U.S. support carbon fee-and-dividend [207];
scription of the climate situation? Does our training, years of college and high school students join in advocacy [208]. A rising
graduate study and decades of experience, not make us well- carbon price creates a level playing field for energy efficiency, re
equipped to advise the public on the climate situation and its pol newable energy, nuclear power, and innovations; it would spur
icy implications? As professionals with deep understanding of the thousands of ‘miracles’ needed for energy transition.
planetary change and as guardians of young people and their fu However, instead, fossil fuels and renewable energy are now sub
ture, do we not have an obligation, analogous to the code of sidized. Thus, nuclear energy has been disadvantaged and ex
ethics of medical professionals, to render to the public our full cluded as a ‘clean development mechanism’ under the Kyoto
and unencumbered diagnosis? That is our objective. Protocol, based on myths about nuclear energy unsupported by

The basis for the following opinions of the first author, to the scientific fact [209]. A rising carbon price is crucial for decarbon
extent not covered in this paper, will be described in a book in ization, but not enough. Long-term planning is needed. Sweden
preparation [2]. We are in the early phase of a climate emer provides an example: 50 years ago, its government decided to re
gency. The present huge planetary energy imbalance assures place fossil fuel power stations with nuclear energy, which led to
that climate will become less tolerable to humanity, with greater its extraordinary and rapid decarbonization (Fig. 32).
climate extremes, before it is feasible to reverse the trend. Second, global cooperation is needed. De facto cooperation be
Reversing the trend is essential—we must cool the planet—for tween the West and China drove down the price of renewable en
the sake of preserving shorelines and saving the world’s coastal ergy. Without greater cooperation, developing nations will be the
cities. Cooling will also address other major problems caused by main source of future GHG emissions (Fig. 28). Carbon-free, dis
global warming. We should aim to return to a climate like that in patchable electricity is a crucial need. Nations with emerging
which civilization developed, in which the nature that we know economies are eager to have modern nuclear power because of
and love thrived. As far as is known, it is still feasible to do that its small environmental footprint. China-U.S. cooperation to de
without passing through irreversible disasters such as many- velop low-cost nuclear power was proposed, but stymied by U.S.
meter sea level rise. prohibition of technology transfer [210]. Competition is normal,
Abundant, affordable, carbon-free energy is essential to but it can be managed if there is a will, reaping benefits of coop
achieve a world with propitious climate, while recognizing the eration over confrontation [211]. Of late, priority has been given
rights and aspirations of all people. The staggering magnitude of instead to economic and military hegemony, despite recognition
the task is implied by global and national carbon intensities: car of the climate threat, and without consultation with young peo
bon emissions per unit energy use (Fig. 32). Global carbon inten ple or seeming consideration of their aspirations. Scientists can
sity must decline to near zero over the next several decades. This support an ecumenical perspective of our shared future by
chart—not vaporous promises of net zero future carbon emis expanding international cooperation. Awareness of the gathering
sions inserted in integrated assessment models—should guide climate storm will grow this decade, so we must increase scien
realistic assessment of progress toward clean energy. Policy must tific understanding worldwide as needed for climate restoration.
include apolitical targeting of support for development of low- Third, we must take action to reduce and reverse Earth’s en
cost carbon-free energy. All nations would do well to study stra ergy imbalance. Highest priority is to phase down emissions, but
tegic decisions of Sweden, which led past decarbonization efforts it is no longer feasible to rapidly restore energy balance via only
(Fig. 32) and is likely to lead in the quest for zero or negative car GHG emission reductions. Additional action is almost surely
bon intensity that will be needed to achieve a bright future for needed to prevent grievous escalation of climate impacts includ
today’s young people and future generations. ing lock-in of sea level rise that could destroy coastal cities
Given the global situation that we have allowed to develop, world-wide. At least several years will be needed to define and
three actions are now essential. gain acceptance of an approach for climate restoration. This ef
First, underlying economic incentives must be installed glob fort should not deter action on mitigation of emissions; on the
ally to promote clean energy and discourage CO2 emissions. contrary, the concept of human intervention in climate is dis
tasteful to many people, so support for GHG emission reductions
1.2 will likely increase. Temporary solar radiation management
Carbon Intensity (MtC/Mtoe) (SRM) will probably be needed, e.g. via purposeful injection of at
1.0 mospheric aerosols. Risks of such intervention must be defined,
Global
India as well as risks of no intervention; thus, the U.S. National
.8
China Academy of Sciences recommends research on SRM [212]. The
Japan
.6 U.S.
Mt. Pinatubo eruption of 1991 is a natural experiment [213, 214]
Germany with a forcing that reached [30] –3 W/m2. Pinatubo deserves a co
.4 France ordinated study with current models. The most innocuous aero
Sweden
sols may be fine salty droplets extracted from the ocean and
.2
sprayed into the air by autonomous sailboats [215]. This ap
0.
proach has been discussed for potential use on a global scale
1970 1980 1990 2000 2010 2020 [216], but it needs research into potential unintended effects
[217]. This decade may be our last chance to develop the knowl
Figure 32. Carbon intensity (carbon emissions per unit energy use) of
several nations and the world. Mtoe ¼ megatons of oil equivalent. Data edge, technical capability, and political will for actions needed to
sources as in Fig. 27. save global coastal regions from long-term inundation.
Politics and climate change while developing the technical knowledge that is needed to navi
Actions needed to drive carbon intensity to zero—most impor gate the stormy sea that their world is setting out upon.
tant a rising carbon fee—are feasible, but not happening. The
first author gained perspective on the reasons why during trips
Acknowledgements
to Washington, DC, and to other nations at the invitation of gov
ernments, environmentalists, and, in one case, oil executives in We thank Eelco Rohling for inviting JEH to describe our perspec
London. Politicians from right (conservative) and left (progres tive on global climate response to human-made forcing. JEH be
sive) parties are affected by fossil fuel interests. The right denies gan to write a review of past work, but a paper on the LGM by
Jessica Tierney et al. [49] and data on changing ship emissions
that fossil fuels cause climate change or says that the effect is ex
provided by Leon Simons led to the need for new analyses and di
aggerated. The left takes up the climate cause but proposes
vision of the paper into two parts. We thank Jessica also for help
actions with only modest effect, such as cap-and-trade with off
ful advice on other related research papers, Jim Zachos and
sets, including giveaways to the fossil fuel industry. The left also
Thomas Westerhold for explanations of their data and interpre
points to work of Amory Lovins as showing that energy efficiency

tations, Ed Dlugokencky of the NOAA Earth System Research
plus renewables (mainly wind and solar energy) are sufficient to
Laboratory for continually updated GHG data, and David Arthur
phase out fossil fuels. Lovins says that nuclear power is not
for pointing out the paper by Steinberger et al. JEH designed the
needed. It is no wonder that the President of Shell Oil would write
study and carried out the research with help of Makiko Sato and
a foreword with praise for Lovins’ book, Reinventing Fire [218], and
Isabelle Sangha; Larissa Nazarenko provided data from GISS
that the oil executives in London did not see Lovins’ work as a models and helped with analysis; Leon Simons provided ship
threat to their business. emission information and aided interpretations; Pushker
Opportunities for progress often occur in conjunction with cri Kharecha provided critical review of the paper; James Zachos
ses. Today, the world faces a crisis—political polarization, espe provided critical interpretation of ocean core data needed for in
cially in the United States—that threatens effective governance. terpretation of Cenozoic climate; Norman Loeb and Karina von
Yet the crisis offers an opportunity for young people to help Schuckmann provided EEI data and insight about implications;
shape the future of the nation and the planet. Ideals professed by Matthew Osman provided paleoclimate data and an insightful re
the United States at the end of World War II were consummated view of an early draft paper; Qinjian Jin provided simulations of
in formation of the United Nations, the World Bank, the Marshall atmospheric sulfate and interpretations; Eunbi Jeong reviewed
Plan, and the Universal Declaration of Human Rights. Progress multiple drafts and advised on presentation; all authors contrib
toward equal rights continued, albeit slowly. The ‘American uted to our research summarized in the paper and reviewed and
dream’ of economic opportunity was real, as most people willing commented on the manuscript. Climate Science, Awareness and
to work hard could afford college. Immigration policy welcomed Solutions, which is directed by JEH and supports MS and PK is a
the brightest; NASA in the 1960s invited scientists from European 501(C3) nonprofit supported 100% by public donations. Principal
countries, Japan, China, India, Canada, and those wanting to stay supporters in the past few years have been the Grantham
found immigration to be straightforward. But the power of spe Foundation, Frank Batten, Eric Lemelson, James and Krisann
cial interests in Washington grew, government became insular Miller, Carl Page, Peter Joseph, Ian Cumming, Gary and Claire
and inefficient, and Congress refused to police itself. Their first Russell, Donald and Jeanne Keith Ferris, Aleksandar Totic, Chris
priority became reelection and maintenance of elite status, sup Arndt, Jeffrey Miller, Morris Bradley and about 150 more contrib
ported by special interests. Thousands of pages of giveaways to utors to annual appeals.
special interests lard every funding bill, including the climate bill
titled ‘Inflation Reduction Act’—Orwellian double-speak—as the
Supplementary data
funding is borrowed from young people via deficit spending. The
public is fed up with the Washington swamp but hamstrung by Supplementary data are available at Oxford Open Climate
Change online.
rigid two-party elections focused on a polarized cultural war.
A political party that takes no money from special interests is
essential to address political polarization, which is necessary if Conflict of interest
the West is to be capable of helping preserve the planet and a
The authors declare that they have no conflict of interest.
bright future for coming generations. Young people showed their
ability to drive an election—via their support of Barack Obama in
2008 and Bernie Sanders in 2016—without any funding from spe Data availability
cial interests. Groundwork is being laid to allow third party can
The data used to create the Figs in this paper are available in the
didates in 2026 and 2028 elections in the U.S. Ranked voting is
Zenodo repository, at https://zenodo.org/record/8419583.
being advocated in every state to avoid the ‘spoiler’ effect of a
third party. It is asking a lot to expect young people to grasp the
situation that they have been handed—but a lot is at stake. As Authors’ contributions
they realize that they are being handed a planet in decline, the James Hansen (Conceptualization [lead], Data curation [equal],
first reaction may be to stamp their feet and demand that gov Formal analysis [lead], Funding acquisition [lead], Investigation
ernments do better, but that has little effect. Nor is it sufficient [lead], Methodology [lead], Project administration [lead],
to parrot big environmental organizations, which are now part of Resources [lead], Software [equal], Supervision [lead], Validation
the problem, as they are partly supported by the fossil fuel indus [lead], Visualization [equal], Writing—original draft [lead],
try and wealthy donors who are comfortable with the status quo. Writing—review and editing [lead]), Makiko Sato (Data curation
Instead, young people have the opportunity to provide the drive [equal], Formal analysis [supporting], Investigation [supporting],
for a revolutionary third party that restores democratic ideals Methodology [supporting], Project administration [supporting],
Resources [supporting], Software [equal], Supervision [support Methodology [supporting], Resources [supporting], Software [sup
ing], Validation [supporting], Visualization [equal], Writing— porting], Supervision [supporting], Validation [supporting],
original draft [supporting], Writing—review and editing [support Writing—review and editing [supporting]).
ing]), Leon Simons (Data curation [supporting], Formal analysis
[supporting], Investigation [supporting], Methodology [support
ing], Resources [supporting], Software [supporting], Supervision Notes
[supporting], Validation [supporting], Visualization [supporting], 1. Drafts of the chapters of Sophie’s Planet relevant to climate sensitivity
are available here; criticisms are welcome.
Writing—review and editing [supporting]), Larissa S. Nazarenko 2. David EE, Jr later became a global warming denier.
(Data curation [supporting], Formal analysis [supporting], 3. GISS (2020) model is described as GISS-E2.1-G-NINT in published papers;
NINT (noninteractive) signifies that the models use specified GHG and
Investigation [supporting], Methodology [supporting], Resources aerosol amounts.
[supporting], Software [supporting], Supervision [supporting], 4. An imbalance of 1 W/m2 for a millennium is enough energy to melt ice
raising sea level 110 m or to raise the temperature of the ocean’s upper
Validation [supporting], Visualization [supporting], Writing— kilometer by 11� C.
review and editing [supporting]), Isabelle Sangha (Formal analysis 5. Tom Delworth (NOAA Geophysical Fluid Dynamics Laboratory), Gokhan

Danabasoglu (National Center for Atmospheric Research), and Jonathan
[supporting], Investigation [supporting], Methodology [supporting],
Gregory (UK Hadley Centre) kindly provided long 2 � CO2 runs of GCMs
Resources [supporting], Software [supporting], Supervision [sup of these leading modeling groups. All three models had response time as
porting], Validation [supporting], Writing—review and editing [sup slow or slower than the GISS GCM.
6. The GISS (2014) model is labeled as GISS-E2-R-NINT and GISS (2020) as
porting]), Pushker Kharecha (Formal analysis [supporting], GISS-E2.1-G-NINT in published papers, where NINT (noninteractive) sig
Investigation [supporting], Methodology [supporting], Resources nifies that the models use specified GHG and aerosol amounts.
7. In Swedish, trapps are stairs. Basalt formations are commonly in layers
[supporting], Software [supporting], Supervision [supporting], from multiple extrusions.
Validation [supporting], Writing—review and editing [supporting]), 8. Small apparent discrepancy is roundoff. CO2 forcing is 9.13 W/m2 and
solar forcing is −1.16 W/m2 at 50 MyBP.
James Zachos (Data curation [supporting], Resources [supporting], 9. Forcing ¼ 4.6 W/m2 assumes that the increase of non-CO2 GHGs is
Writing—review and editing [supporting]), Karina von human-made. This is true for CFCs and most trace gases, but a small
Schuckmann (Formal analysis [supporting], Investigation [support part of CH4 and N2O growth could be a slow feedback, slightly reducing
the GHG forcing.
ing], Methodology [supporting], Resources [supporting], Software 10. 9.9� C for ECS ¼ 1.2� C per W/m2; 10.1� C for ECS ¼ 1.22� C per W/m2 (the
[supporting], Supervision [supporting], Validation [supporting], precise ECS for 7� C LGM cooling).
11. Two significant flaws in the derivation of this ‘alternative aerosol sce
Writing—review and editing [supporting]), Norman G. Loeb (Data nario’ were largely offsetting: (1) the intermediate climate response
curation [supporting], Formal analysis [supporting], Investigation function employed was too ‘fast’, but (2) this was compensated by use of
a low climate sensitivity of 3� C for 2 � CO2.
[supporting], Methodology [supporting], Resources [supporting], 12. In the absence of a response function from a GCM with ECS ¼ 4� C, we
Software [supporting], Supervision [supporting], Validation [sup use the normalized response function of the GISS (2020) model and put k
¼ 1� C per W/m2 in Equation (5).
porting], Visualization [supporting], Writing—review and editing 13. Jay Zwally, Eric Rignot, Konrad Steffen, and Roger Braithwaite.
[supporting]), Matthew B. Osman (Data curation [supporting],
Formal analysis [supporting], Investigation [supporting],
Methodology [supporting], Resources [supporting], Software [sup
porting], Supervision [supporting], Validation [supporting], References
Visualization [supporting], Writing—review and editing [support
1.00 Tyndall J. On the absorption and radiation of heat by gases and
ing)], Qinjian Jin (Data curation [supporting], Formal analysis [sup
porting], Investigation [supporting], Methodology [supporting], vapours. Phil Mag 1861;22:169–194, 273–285.
Resources [supporting], Software [supporting], Supervision [sup 2.00 Hansen J. Greenhouse Giants, Chapter 15 in Sophie’s Planet. New
porting], Validation [supporting], Visualization [supporting], York: Bloomsbury, 2024.
Writing—review and editing [supporting]), George Tselioudis 3.00 Revelle R, Broecker W, Craig H et al. Appendix Y4 atmospheric
(Formal analysis [supporting], Investigation [supporting], carbon dioxide. In: Hornig J, York HF, Branscomb LM et al.
Methodology [supporting], Resources [supporting], Software [sup (eds,) President’s Science Advisory Committee. Restoring the Quality
porting], Supervision [supporting], Writing—review and editing of Our Environment. Washington: The White House,
[supporting]), Eunbi Jeong (Formal analysis [supporting], 1965, 111–33.
Investigation [supporting], Methodology [supporting], Resources 4.00 Charney J, Arakawa A, Baker D et al. Carbon Dioxide and Climate:
[supporting], Software [supporting], Supervision [supporting], A Scientific Assessment. Washington: National Academy of
Validation [supporting], Writing—review and editing [supporting]), Sciences Press, 1979.
Andrew Lacis (Formal analysis [supporting], Investigation [sup 5.00 Nierenberg WA. Changing Climate: Report of the Carbon Dioxide
porting], Methodology [supporting], Resources [supporting], Assessment Committee. Washington: National Academies
Software [supporting], Supervision [supporting], Validation [sup Press, 1983.
porting], Writing—review and editing [supporting]), Reto Ruedy 6.00 Hansen JE, Takahashi T (eds). AGU Geophysical Monograph 29
(Formal analysis [supporting], Investigation [supporting], Climate Processes and Climate Sensitivity. Washington: American
Methodology [supporting], Resources [supporting], Software [sup Geophysical Union, 1984.
porting], Supervision [supporting], Validation [supporting], 7.00 Hansen J, Lacis A, Rind D et al. Climate sensitivity: analysis of
Writing—review and editing [supporting]), Gary Russell (Formal feedback mechanisms. In: Hansen JE, Takahashi T (eds), AGU
analysis [supporting], Investigation [supporting], Methodology Geophysical Monograph 29 Climate Processes and Climate
[supporting], Resources [supporting], Software [supporting], Sensitivity. Washington: American Geophysical Union,
Supervision [supporting], Validation [supporting], Writing—review 1984, 130–63.
and editing [supporting]), Junji Cao (Formal analysis [supporting], 8.00 David EE Jr. Inventing the future: energy and the CO2
Investigation [supporting], Methodology [supporting], Resources “Greenhouse Effect”. In: Hansen JE, Takahashi T (eds). AGU
[supporting], Software [supporting], Supervision [supporting], Geophysical Monograph 29 Climate Processes and Climate
Validation [supporting], Writing—review and editing [supporting]), Sensitivity. Washington: American Geophysical Union,
Jing Li (Formal analysis [supporting], Investigation [supporting], 1984, 1–5.
9.00 Oreskes N, Conway E. Merchants of Doubt: How a Handful of (eds), Climate Change 2013: The Physical Science Basis. Contribution
Scientists Obscured the Truth on Issues from Tobacco Smoke to Global of Working Group I to the Fifth Assessment Report of the
Warming. London: Bloomsbury, 2010. Intergovernmental Panel on Climate Change. Cambridge and New
10.0 Intergovernmental Panel on Climate Change. History of the York: Cambridge University Press, 2013.
IPCC. https://www.ipcc.ch/about/history (date last accessed 30.0 Hansen J, Sato M, Ruedy R et al. Efficacy of climate forcings.
7 March 2023). J Geophys Res 2005;110:D18104.
11.0 United Nations Framework Convention on Climate Change. 31.0 Lohmann U, Rotstayn L, Storelvmo T et al. Total aerosol effect:
What is the United Nations Framework Convention on Climate radiative forcing or radiative flux perturbation? Atmos Chem
Change? https://unfccc.int/process-and-meetings/what-is-the- Phys 2010;10:3235–46.
united-nations-framework-convention-on-climate-change) 32.0 Kelley M, Schmidt GA, Nazarenko L et al. GISS-E2.1: configura
(date last accessed 30 November 2022).
tions and climatology. J Adv Model Earth Syst 2020;12:
12.0 IPCC. Climate Change 2021: The Physical Science Basis. [Masson-
e2019MS002025.
Delmotte V, Zhai P, Pirani A et al. (eds)], Cambridge and New
33.0 Miller RL, Schmidt GA, Nazarenko L et al. CMIP6 historical sim

York: Cambridge University Press, 2021.
ulations (1850-2014) with GISS-E2.1. J Adv Model Earth Syst
13.0 Hansen J, Sato M, Hearty P et al. Ice melt, sea level rise and
2021;13:e2019MS002034.
superstorms: evidence from paleoclimate data, climate
34.0 Eyring V, Bony S, Meehl GA et al. Overview of the Coupled
modeling, and modern observations that 2 C global warming
Model Intercomparison Project Phase 6 (CMIP6) experimental
could be dangerous. Atmos Chem Phys 2016;16:3761–812.
design and organization. Geosci Model Dev 2016;9:1937–58.
14.0 Hansen J. Foreword: uncensored science is crucial for global
35.0 Lacis AA, Oinas V. A description of the correlated k distributed
conservation. In: DellaSala DA (ed), Conservation Science and
method for modeling nongray gaseous absorption, thermal
Advocacy for a Planet in Peril. Amsterdam: Elsevier, 2021, 451.
15.0 Bode HW. Network Analysis and Feedback Amplifier Design. New emission, and multiple scattering in vertically inhomogeneous
York: Van Nostrand, 1945. atmospheres. J Geophys Res 1991;96:9027–63.
16.0 Lacis A, Hansen J, Lee P et al. Greenhouse effect of trace gases, 36.0 Rothman L, Rinsland C, Goldman A et al. The HITRAN molecu
1970-1980. Geophys Res Lett 1981;8:1035–8. lar spectroscopic database and HAWKS (HITRAN Atmospheric
17.0 CLIMAP Project Members. Seasonal reconstruction of the Workshation) 1996 edition. J Quan Spec Rad Trans 1998;
Earth’s surface at the last glacial maximum. Geol Soc Amer, 60:665–710.
Map and Chart Series, No. 36, Geological Society of 37.0 Prather M, Ehhalt D. Chapter 4: Atmospheric chemistry and
America 1981. greenhouse gases. In: Houghton JT (ed), Climate Change 2001:
18.0 Manabe S, Stouffer RJ. Sensitivity of a global climate model to The Scientific Basis. New York: Cambridge University,
an increase of CO2 concentration in the atmosphere. J Geophys 2001, 239–87.
Res 1980;85:5529–54. 38.0 Hansen J, Sato M. Greenhouse gas growth rates. Proc Natl Acad
19.0 Manabe S. Carbon dioxide and climate change. Adv Geophys Sci U S A 2004;101:16109–14.
1983;25:39–82. 39.0 Columbia University. MPTG and OTG data: www.columbia.edu/
20.0 Klein SA, Hall A, Norris JR et al. Low-cloud feedbacks from ~mhs119/GHGs/TG_F.1900-1990.txt and http://www.colum
cloud-controlling factors: a review. Surv Geophys 2017; bia.edu/~mhs119/GHGs/TG_F.1992-2021.txt (date last accessed
38:1307–29. 9 August 2023).
21.0 Sherwood SC, Webb MJ, Annan JD et al. An assessment of 40.0 Jouzel J, Masson-Delmotte V, Cattani O et al. Orbital and mil
Earth’s climate sensitivity using multiple lines of evidence. Rev lennial Antarctic climate variability over the past 800,000
Geophys 2020;58:e2019RG000678.
years. Science 2007;317:793–6.
22.0 Zelinka MD, Zhou C, Klein SA. Insights from a refined decom
41.0 Luthi D, Le Floch M, Bereiter B et al. High-resolution carbon di
position of cloud feedbacks. Geophys Res Lett 2016;43:9259–69.
oxide concentration record 650,000-800,000 years before pre
23.0 Zelinka M, Tan I, Oreopoulos L et al. Detailing cloud property
sent. Nature 2008;453:379–82.
feedbacks with a regime-based decomposition. Clim Dyn 2023;
42.0 Hays JD, Imbrie J, Shackleton NJ. Variation in the Earth’s orbit:
60:2983–3003.
pacemaker of the ice ages. Science 1976;194:1121–32.
24.0 Rind D, Peteet D. Terrestrial conditions at the last glacial maxi
43.0 Lorius C, Jouzel J, Raynaud D et al. The ice-core record: climate
mum and CLIMAP sea-surface temperature estimates: Are
sensitivity and future greenhouse warming. Nature 1990;
they consistent? Quat Res 1985;24:1–22.
347:139–45.
25.0 Rohling EJ, Marino G, Foster GL et al. Comparing climate sensi
44.0 Zachos J, Pagani M, Sloan L et al. Trends, rhythms, and aberra
tivity, past and present. Ann Rev Mar Sci 2018;10:261–88.
26.0 IPCC. Climate Change 2014: Synthesis Report. Contribution of tions in global climate 65 Ma to present. Science 2001;
Working Groups I, II and III to the Fifth Assessment Report of the 292:686–93.
Intergovernmental Panel on Climate Change. [Core Writing Team, 45.0 Hansen J, Sato M, Kharecha P et al. Climate change and trace
Pachauri RK, Meyer LA (eds)]. Geneva: IPCC, 2014. gases. Philos Trans A Math Phys Eng Sci 2007;365:1925–54.
27.0 Andrews T, Gregory JM, Paynter D et al. Accounting for chang 46.0 Ruddiman WF, Fuller DQ, Kutzbach JE et al. Late Holocene cli
ing temperature patterns increases historical estimates of cli mate: natural or anthropogenic? Rev Geophys 2016;54:93–118.
mate sensitivity. Geophys Res Lett 2018;45:8490–9. 47.0 Schilt A, Baumgartner M, Schwander J et al. Atmospheric ni
28.0 Rugenstein M, Bloch-Johnson J, Abe-Ouchi A et al. trous oxide during the last 140,000 years. Earth Planet Sci Lett
LongRunMIP: motivation and design for a large collection of 2010;300:33–43.
millennial-length AOGCM simulations. Bull Amer Meteorol Soc 48.0 Hansen J, Nazarenko L, Ruedy R et al. Earth's energy imbalance:
2019;100:2551–70. Confirmation and implications. Science 2005;308:1431–5.
29.0 Myhre G, Shindell D, Bre �on F-M et al. Anthropogenic and natu 49.0 Tierney JE, Zhu J, King J et al. Glacial cooling and climate sensi
ral radiative forcing. In: Stocker TF, Qin D, Plattner G-K et al. tivity revisited. Nature 2020;584:569–73.
50.0 Osman MB, Tierney JE, Zhu J et al. Globally resolved surface 71.0 Lunt DJ, Haywood AM, Schmidt GA et al. Earth system sensitiv
temperatures since the Last Glacial Maximum. Nature 2021; ity inferred from Pliocene modelling and data. Nature Geosci
599:239–44. 2010;3:60–4.
51.0 Seltzer AM, Ng J, Aeschbach W et al. Widespread six degrees 72.0 Beerling DJ, Fox A, Stevenson DS et al. Enhanced chemistry-
Celsius cooling on land during the Last Glacial Maximum. climate feedbacks in past greenhouse worlds. Proc Natl Acad Sci
Nature 2021;593:228–32. U S A 2011;108:9770–5.
52.0 Schneider T, Teixeira J, Bretherton CS et al. Climate goals and 73.0 Bryan K, Komro FG, Manabe S et al. Transient climate response
computing the future of clouds. Nature Clim Change 2017;7:3–5. to increasing atmospheric carbon dioxide. Science 1982;
53.0 Pincus R, Forster PM, Stevens B. The radiative forcing model 215:56–8.
intercomparison project (RFMIP): experimental protocol for 74.0 Hansen J, Russell G, Lacis A et al. Climate response times: de
CMIP6. Geosci Model Dev 2016;9:3447–60. pendence on climate sensitivity and ocean mixing. Science
54.0 Kageyama M, Braconnot P, Harrison SP et al. The PMIP4 contri 1985;229:857–9.
bution to CMIP6 – Part 1: overview and over-arching analysis 75.0 Hansen J. Climate Threat to the Planet, American Geophysical

plan. Geosci Model Dev 2018;11:1033–57. Union, San Francisco, California, 17 December 2008. http://
55.0 Hegerl GC, Zwiers FW, Braconnot P et al. Chapter 9: www.columbia.edu/~jeh1/2008/AGUBjerknes20081217.pdf.
Understanding and attributing climate change. In: Solomon (date last accessed 3 December 2022).
SD (ed), Climate Change 2007: The Physical Science Basis. New 76.0 Good P, Gregory JM, Lowe JA. A step-response simple climate
York: Cambridge University, 2007, 663–745 model to reconstruct and interpret AOGCM projections.
56.0 Yoshimori M, Yokohata T, Abe-Ouchi A. A comparison of cli Geophys Res Lett 2011;38:e2010GL0452008.
mate feedback strength between CO2 doubling and LGM 77.0 Schmidt GA, Kelley M, Nazarenko L et al. Configuration and as
experiments. J Clim 2009;22:3374–95.
sessment of the GISS ModelE2 contributions to the CMIP5 ar
57.0 Stap LB, Kohler P, Lohmann G. Including the efficacy of land
chive. J Adv Model Earth Syst 2014;6:141–84.
ice changes in deriving climate sensitivity from paleodata.
78.0 Prather MJ. Numerical advection by conservation of second or
Earth Syst Dynam 2019;10:333–45.
der moments. J Geophys Res 1986;91:6671–81.
58.0 Koppen W. Das geographische system der climate. In: Koppen
79.0 Romanou A, Marshall J, Kelley M et al. Role of the ocean’s
W, Geiger G (eds), Handbuch Der Klimatologie 1(C). Berlin:
AMOC in setting the uptake efficiency of transient tracers.
Boentraeger, 1936.
Geophys Res Lett 2017;44:5590–8.
59.0 Kohler P, Bintanja R, Fischer H et al. What caused Earth’s tem
80.0 von Schuckmann K, Cheng L, Palmer MD et al. Heat stored in
perature variations during the last 800,000 years? Data-based
the Earth system: where does the energy go? Earth Syst Sci Data
evidence on radiative forcing and constraints on climate sensi
2020;12:2013–41.
tivity. Quat Sci Rev 2010;29:129–45.
81.0 Loeb NG, Johnson GC, Thorsen TJ et al. Satellite and ocean data
60.0 Hansen J, Sato M, Kharecha P et al. Target atmospheric CO2:
reveal marked increase in Earth’s heating rate. Geophys Res Lett
where should humanity aim? Open Atmos Sci J 2008;2:217–31.
2021;48:e2021GL093047.
61.0 Rabineau M, Berne � S, Olivet J-L et al. Paleo sea levels reconsid
82.0 Hansen J, Johnson D, Lacis A et al. Climate impact of increasing
ered from direct observation of paleoshoreline position during
atmospheric carbon dioxide. Science 1981;213:957–66.
Glacial Maxima (for the last 500,000 yr). Earth Planet Sci Lett
83.0 Kamae Y, Watanabe M, Ogura T et al. Rapid adjustments of
2006;252:119–37.
cloud and hydrological cycle to increasing CO2: a review. Curr
62.0 Rohling EJ, Hibbert FD, Williams FH et al. Differences between
Clim Change Rep 2015;1:103–13.
the last two glacial maxima and implications for ice-sheet,
� 18O, and sea-level reconstructions. Quat Sci Rev 2017; 84.0 Zelinka MD, Myers TA, McCoy DT et al. Causes of higher cli
o
mate sensitivity in CMIP6 models. Geophys Res Lett 2020;47:
176:1–28.
e2019GL085782.
63.0 Hansen J, Sato M, Kharecha P et al. Young people’s burden: re
85.0 Crowley TJ. Pliocene climates: the nature of the problem.
quirement of negative CO2 emissions. Earth Syst Dynam 2017;
8:577–616. Marine Micropaleontol 1996;27:3–12.
64.0 Hoffman JS, Clark PU, Parnell AC et al. Regional and global sea- 86.0 DeConto RM, Pollard D. Rapid Cenozoic glaciation of
surface temperatures during the last interglaciation. Science Antarctica induced by declining atmospheric CO2. Nature 2003;
2017;355:276–9. 421:245–9.
65.0 Ruth U, Barnola JM, Beer J et al. EDML1: a chronology for the 87.0 Lacis AA, Schmidt GA, Rind D et al. Atmospheric CO2: principal
EPICA deep ice core from Dronning Maud Land, Antarctica, control knob governing Earth’s temperature. Science 2010;
over the last 150 000 years. Clim Past 2007;3:475–84. 330:356–9.
66.0 Hansen J, Sato M, Russell G et al. Climate sensitivity, sea level, 88.0 Rae JWB, Zhang YG, Liu X et al. Atmospheric CO2 over the past
and atmospheric carbon dioxide. Philos Trans A Math Phys Eng 66 million years from marine archives. Annu Rev Earth Planet Sci
Sci 2013;371:20120294. 2021;49:609–41.
67.0 Russell GL, Miller JR, Rind D. A coupled atmosphere-ocean 89.0 Steinthorsdottir M, Vajda V, Pole M et al. Moderate levels of
model for transient climate change studies. Atmos Ocean 1995; Eocene pCO2 indicated by Southern Hemisphere fossil plant
33:683–730. stomata. Geology 2019;47:914–8.
68.0 Hoffman PF, Schrag DP. The snowball Earth hypothesis: testing 90.0 Westerhold T, Marwan N, Drury AJ et al. An astronomically
the limits of global change. Terra Nova 2002;14:129–55. dated record of Earth’s climate and its predictability over the
69.0 Sackmann J, Boothroyd AI, Kraemer KE. Our Sun. III. Present last 66 million years. Science 2020;369:1383–7.
and future. Astrophys J 1993;418:457–68. 91.0 Yatheesh V, Dyment J, Bhattacharya GC et al. Detailed struc
70.0 Meraner K, Mauritsen T, Voigt A. Robust increase in equilib ture and plate reconstructions of the central Indian Ocean be
rium climate sensitivity under global warming. Geophys Res tween 83.0 and 42.5 Ma (chrons 34 and 20). J Geophys Res: Solid
Lett 2013;40:5944–8. Earth 2019;124:4305–22.
92.0 Cutler KB, Edwards RL, Taylor FW et al. Rapid sea-level fall and 114. Storey M, Duncan RA, Tegner C. Timing and duration of volca
deep-ocean temperature change since the last interglacial pe nism in the North Atlantic Igneous Province: implications for
riod. Earth Planet Sci Lett 2003;206:253–71. geodynamics and links to the Iceland hotspot. Chem Geol 2007;
93.0 Siddall M, Honisch B, Waelbroeck C et al. Changes in deep 241:264–81.
Pacific temperature during the mid-Pleistocene transition and 115. Svensen H, Planke S, Malthe-Sorenssen A et al. Release of
Quaternary. Quatern Sci Rev 2010;29:170–81. methane from a volcanic basin as a mechanism for initial
94.0 Rohling EJ, Grant K, Bolshaw M et al. Antarctic temperature Eocene global warming. Nature 2004;429:542–5.
and global sea level closely coupled over the past five glacial 116. Gutjahr M, Ridgwell A, Sexton PF et al. Very large release of
cycles. Nature Geosci 2009;2:500–4. mostly volcanic carbon during the Palaeocene Thermal
95.0 Seltzer AM, Blard P-H, Sherwood SC et al. Terrestrial amplifica Maximum. Nature 2017;548:573–7.
tion of past, present, and future climate change. Sci Adv 2023; 117. Frieling J, Peterse F, Lunt DJ et al. Widespread warming before
9:eadf8119. and elevated barium burial during the Paleocene-Eocene ther
96.0 Zhu J, Poulsen CJ, Tierney JE. Simulation of Eocene extreme mal maximum: evidence for methane hydrate release?

warmth and high climate sensitivity through cloud feedbacks. Paleoceanogr Paleoclimatol 2019;34:546–66.
Sci Adv 2019;5:eaax1874. 118. Berndt C, Planke S, Alvarez Zarikian CA et al. Shallow-water
97.0 Scotese C. PALEOMAP PaleoAtlas for GPlates, 2016. https://www. hydrothermal venting linked to the Palaeocene-Eocene
earthbyte.org/paleomap-paleoatlas-for-gplates/ (March 2023, Thermal Maximum. Nature Geosci 2023;16:803–9.
date last accessed). 119. Walker JCG, Hays PB, Kasting JF. A negative feedback mecha
98.0 Hansen J. Storms of My Grandchildren. New York: Bloomsbury, 2009. nism for the long-term stabilization of Earth’s surface temper
99.0 Berner RA. The Phanerozoic Carbon Cycle: CO2 and O2. New York: ature. J Geophys Res 1981;86:9776–82.
Oxford University Press, 2004. 120. Steinberger B, Spakman W, Japsen P et al. The key role of global
100. Rohling EJ. The Climate Question: natural Cycles, Human Impact, solid-Earth processes in preconditioning Greenland’s glacia
Future Outlook. New York: Oxford University Press, 2019. tion since the Pliocene. Terra Nova 2015;27:1–8.
101. Merdith AS, Williams SE, Brune S et al. Rift and plate boundary 121. Foster GL, Hull P, Lunt DJ et al. Placing our current
evolution across two supercontinent cycles. Global Plan Chan ‘hyperthermal’ in the context of rapid climate change in our
2019;173:1–14. geological past. Phil Trans Roy Soc A 2018;376:200170086.
102. Peace AL, Phethean JJJ, Franke D et al. A review of Pangea dis 122. Tierney JE, Zhu J, Li M. Spatial patterns of climate change
persal and large igneous provinces – in search of a causative across the Paleocene-Eocene thermal maximum. Proc Natl
mechanism. Earth-Science Rev 2020;206:102902. Acad Sci 2022;119:e2205326119.
103. Baksi AK. Comment on “40Ar/39Ar dating of the Rajahmundry 123. Hopcroft PO, Ramstein G, Pugh TAM et al. Polar amplification
Traps, eastern India and their relationship to the Deccan of Pliocene climate by elevated trace gas radiative forcing. Proc
Traps” by Knight et al. [Earth Planet Sci. Lett. 208 (2003) 85-99]. Natl Acad Sci U S A 2020;117:23401–7.
Earth Planet Sci Lett 2005;239:368–73. 124. Schaller MF, Fung MK. The extraterrestrial impact evidence at
104. Guo Z, Wilson M, Dingwell D et al. India-Asia collision as a the Palaeocene-Eocene boundary and sequence of environ
driver of atmospheric CO2 in the Cenozoic. Nature Comm 2021; mental change on the continental shelf. Phil Trans R Soc A 2018;
12:3891. 376:20170081.
105. Raymo ME, Ruddiman WF. Tectonic forcing of late Cenozoic 125. Turner SK. Constraints on the onset duration of the Paleocene-
climate. Nature 1992;359:117–22. Eocene Thermal Maximum. Phil Trans R Soc A 2018;
106. Ramos EJ, Lackey JS, Barnes JD et al. Remnants and rates of 376:20170082.
metamorphic decarbonation in continental arcs. Gsat 2020; 126. Zachos JC, McCarren H, Murphy B et al. Tempo and scale of late
30:4–10. Paleocene and early Eocene carbon isotope cycles: implica
107. Bufe A, Hovius N, Emberson R et al. Co-variation of silicate, car tions for the origin of hyperthermals. Earth Plan Sci Lett 2010;
bonate and sulfide weathering drives CO2 release with erosion. 299:242–9.
Nat Geosci 2021;14:211–6. 127. Nichols JE, Peteet DM. Rapid expansion of northern peatlands
108. Lee CTA, Shen B, Slotnick BS et al. Continental arc-island arc and doubled estimate of carbon storage. Nat Geosci 2019;
fluctuations, growth of crustal carbonates, and long-term cli 12:917–21.
mate change. Geosphere 2013;9:21–36. 128. Hanson PJ, Griffiths NA, Iverson CM et al. Rapid net carbon loss
109. McKenzie NR, Horton BK, Loomis SE et al. Continental arc vol from a whole-ecosystem warmed peatland. AGU Advan 2020;
canism as the principal driver of icehouse-greenhouse vari 1:e2020AV000163.
ability. Science 2016;352:444–7. 129. Bowen GJ, Maibauer BJ, Kraus MJ et al. Two massive, rapid
110. Petersen KD, Schiffer C, Nagel T. LIP formation and protracted releases of carbon during the onset of the Palaeocene-Eocene
lower mantle upwelling induced by rifling and delamination. thermal maximum. Nature Geosci 2015;8:44–7.
Scientific Rep 2018;8:16578. 130. Archer D, Buffett B, Brovkin V. Ocean methane hydrates as a
111. Eldholm E, Grue K. North Atlantic volcanic margins: dimen slow tipping point in the global carbon cycle. Proc Natl Acad Sci
sions and production rates. J Geophys Res 1994;99:2955–68. U S A 2009;106:20596–601.
112. Ji S, Nie J, Lechler A et al. A symmetrical CO2 peak and asym 131. Archer D, Eby M, Brovkin V et al. Atmospheric lifetime of fossil
metrical climate change during the middle Miocene. Earth Plan fuel carbon dioxide. Annu Rev Earth Planet Sci 2009;37:117–34.
Sci Lett 2018;499:134–44. 132. Nunes F, Norris RD. Abrupt reversal in ocean overturning dur
113. Babila TL, Foster GL. The Monterey Event and the Paleocene- ing the Palaeocene/Eocene warm period. Nature 2006;439:60–3.
Eocene Thermal Maximum: two contrasting oceanic carbonate 133. World Health Organization. Ambient (Outdoor) Air Pollution.
system responses to LIP emplacement and eruption. In: https://www.who.int/en/news-room/fact-sheets/detail/ambi
Dickson A, Bekker A (eds), AGU Geographical Monograph 255, ent-(outdoor)-air-quality-and-health (date last accessed 23
Washington: American Geophysical Union, 2021, 403–416. June 2022).
134. Vohra K, Vodonos A, Schwartz J et al. Global mortality from out 156. Lenssen NJL, Schmidt GA, Hansen JE et al. Improvements in the
door fine particle pollution generated by fossil fuel combustion: GISTEMP uncertainty model. JGR Atmospheres 2019;124:6307–26.
results from GEOS-Chem. Environ Res 2021;195:110754. 157. Jin Q, Grandey BS, Rothenberg D et al. Impacts on cloud radia
135. Marcott SA, Shakun JD, Clark PU et al. A reconstruction of re tive effects induced by coexisting aerosols converted from in
gional and global temperature for the last 11,300. Science 2013; ternational shipping and maritime DMS emissions. Atmos
339:1198–201. Chem Phys 2018;18:16793–808.
136. Tardif R, Hakim GJ, Perkins WA et al. Last Millenium Reanalysis 158. Hansen J, Rossow W, Carlson B et al. Low-cost long-term moni
with an expanded proxy database and seasonal proxy model toring of global climate forcings and feedbacks. Clim Chan
ing. Clim Past 2019;15:1251–73. 1995;31:247–71.
137. Watson AJ, Garabato ACN. The role of Southern Ocean mixing 159. Glassmeier F, Hoffmann F, Johnson JS et al. Aerosol-cloud-cli
and upwelling in glacial-interglacial atmospheric CO2 change. mate cooling overestimated by ship-track data. Science 2021;
Tellus 2006;58:73–87. 371:485–9.
138. Wikipedia. File:Post-Glacial Sea Level.png. https://commons.wiki 160. Manshausen P, Watson-Parris D, Christensen MW et al.

media.org/wiki/File:Post-Glacial_Sea_Level.png (date last Invisible ship tracks show large cloud sensitivity to aerosol.
accessed 3 December 2022). Nature 2022;610:101–6.
139. Barber B. Resistance by scientists to scientific discovery. Science 161. Wall CJ, Norris JR, Possner A et al. Assessing effective radiative
1961;134:596–602. forcing from aerosol-cloud interactions over the global ocean.
140. Hoffman PF, Kaufman AJ, Halverson GP et al. A Neoproterozoic Proc Natl Acad Sci U S A 2022;119:e2210481119.
Snowball Earth. Science 1998;281:1342–6. 162. Forster P, Storelvmo T, Armour K et al. The Earth’s energy budget,
141. Alvarez L, Alvarez W, Asaro F et al. Extraterrestrial Cause for climate feedbacks, and climate sensitivity. In: Masson-Delmotte
the Cretaceous-Tertiary Extinction. Science 1980;208:1095–108. V (ed), Climate Change 2021: The Physical Science Basis. New York,
142. Mishchenko MI, Cairns B, Kopp G et al. Accurate monitoring of Cambridge: Cambridge University Press, 2021, 923–1054.
terrestrial aerosols and total solar irradiance: Introducing the 163. International Maritime Organization (IMO), MEPC.176(58).
Glory mission. Bull Amer Meteorol Soc 2007;88:677–92. Amendments to the annex of the protocol of 1997 to amend
143. Hansen J, Rossow W, Fung I. Long-Term Monitoring of Global the international convention for the prevention of pollution
Climate Forcings and Feedbacks. Washington: NASA Conference from ships, 1973, as modified by the protocol of 1978 relating
Publication 3234, 1993. thereto (Revised MARPOL, Annex VI), 2008. https://wwwcdn.
144. Bellouin N, Quaas J, Gryspeerdt E et al. Bounding global aerosol imo.org/localresources/en/OurWork/Environment/
radiative forcing of climate change. Rev Geophys 2020;58: Documents/176(58).pdf (February 2013, date last accessed).
e2019RG000660. 164. Gryspeerdt E, Smith TWP, O'Keeffe E et al. The impact of ship
145. Kruzman D. Wood-burning stoves raise new health concerns. emission controls recorded by cloud properties. Geophys Res
Undark Magazine 2022, 2 March (6 February 2023, date Lett 2019;46:547–55.
last accessed). 165. International Maritime Organization. IMO 2020–Cutting Sulphur
146. Glojek K, Mocnik G, Alas HDC et al. The impact of temperature Oxide Emissions, Lowers Limit on Sulfur Content of Marine Fuels
inversions on black carbon and particle mass concentrations from 3.5% to 0.5%. https://www.imo.org/en/MediaCentre/
in a mountainous area. Atmos Chem Phys 2022;22:5577–601. HotTopics/Pages/Sulphur-2020.aspx (date last accessed 5
147. Rutgard O. Why is Britain Taking the Axe to Wood-Burning Stoves? December 2022).
Bloomberg Green, 4 February 2023. 166. Yuan T, Song H, Wood R et al. Global reduction in ship-tracks
148. Day JW, Gunn JD, Folan WJ et al. Emergence of complex socie from sulfur regulations for shipping fuel. Sci Adv 2022;
ties after sea level stabilized. EOS Transactions 2007;88:169–70. 8:eabn7988.
149. VanCuren RA. Asian aerosols in North America: extracting the 167. Columbia University. Data sources, graphs. http://www.colum
chemical composition and mass concentration of the Asian bia.edu/~mhs119/Solar/ (date last accessed 23 October 2022).
continental aerosol plume from long-term aerosol records in 168. Loeb NG, Thorsen TJ, Rose FG et al. Recent variations in EEI,
the western United States. J Geophs Res Atmos 2003;108: SST & clouds. ERB Workshop, Hamburg, Germany, 12–14
D20, 4623. October, 2022 (date last accessed 3 December 2022).
150. Knutti R. Why are climate models reproducing the observed 169. Sato M. Sea ice area. Columbia University webpage (date last
global surface warming so well? Geophys Res Lett 2008; accessed 5 November 2022).
35:L18704. 170. McCoy DT, Burrows SM, Wood R et al. Natural aerosols explain
151. Hansen J, Sato M, Kharecha P et al. Earth's energy imbalance seasonal and spatial patterns of Southern Ocean cloud albedo.
and implications. Atmos Chem Phys 2011;11:13421–49. Sci Adv 2015;1:e1500157.
152. Koch D, Bauer SE, Del Genio A et al. Coupled aerosol- 171. Dunne JP, Winton M, Bacmeister J et al. Comparison of equilib
chemistry-climate twentieth-century model investigation: rium climate sensitivity estimates from slab ocean, 150-year,
trends in short-lived species and climate responses. J Clim and longer simulations. Geo Res Lett 2020;47:e2020GL088852.
2011;24:2693–714. 172. Forster PM, Maycock AC, McKenna CM et al. Latest climate
153. Novakov T, Ramanathan V, Hansen JE et al. Large historical models confirm need for urgent mitigation. Nat Clim Chang
changes of fossil-fuel black carbon aerosols. Geophys Res Lett 2020;10:7–10.
2003;30:1324. 173. Liu Z, Zhu J, Rosenthal Y et al. The Holocene temperature co
154. Bauer SE, Tsigaridis K, Faluvegi G et al. Historical (1850-2014) nundrum. Proc Natl Acad Sci U S A 2014;111:E3501–E3505.
aerosol evolution and role on climate forcing using the GISS 174. Quaas J, Jia H, Smith C et al. Robust evidence for reversal of the
ModelE2.1 contribution to CMIP6. J Adv Model Earth Syst 2020; trend in aerosol effective climate forcing. Atmos Chem Phys
12:e2019MS001978. 2022;22:12221–39.
155. Hansen J, Ruedy R, Sato M et al. Global surface temperature 175. Bauer SE, Tsigaridis K, Faluvegi G et al. The turning point of the
change. Rev Geophys 2010;48:RG4004. aerosol era. J Adv Mod Earth Syst 2022;14:e2022MS003070.
176. Diamond MS. Detection of large-scale cloud microphysical 196. Rignot E, Mouginot J, Scheuchl B et al. Four decades of
changes and evidence for decreasing cloud brightness within a Antarctic ice shelf mass balance from 1979-2017. Proc Natl
major shipping corridor after implementation of the Acad Sci U S A 2018;116:1096–103.
International Maritime Organization 2020 fuel sulfur regula 197. Otosaka IN, Shepherd A, Ivins ER et al. Mass balance of the
tions. Atmos Chem Phys 2023;23:8259–69. Greenland and Antarctic ice sheets from 1992 to 2020. Earth
177. Hefner M, Marland G, Boden T et al. Global, Regional, and National Syst Sci Data 2023;15:1597–616.
Fossil-Fuel CO2 Emissions. Research Institute for Environment, 198. Hansen J, Sato M, Ruedy R et al. Dangerous human-made inter
Energy, and Economics, Appalachian State University, Boone, ference with climate: A GISS modelE study. Atmos Chem Phys
NC, USA. https://energy.appstate.edu/cdiac-appstate/data- 2007;7:2287–312.
products (date last accessed 20 August 2023). 199. Matthews HD, Gillett NP, Stott PA et al. The proportionality of
178. Energy Institute. 2023 Statistical Review of World Energy (date
global warming to cumulative carbon emissions. Nature 2009;
last accessed 20 August 2023).
459:829–32.
179. Hansen J, Sato M, Ruedy R et al. Global warming in the twenty-
200. Hansen J, Sato M. Regional Climate Change and National

first century: an alternative scenario. Proc Natl Acad Sci U S A
Responsibilities. Environ Res Lett 2016;11:034009.
2000;97:9875–80.
201. Paris Agreement 2015. UNFCCC Secretariat, 2015. https://
180. Hansen JE (ed.). Air Pollution as a Climate Forcing: A Workshop.
www.google.com/aclk?sa=l&ai=DChcSEwjrgqiAkf2BAxUz2Ew
NASA Goddard Institute for Space Studies, 2002.
CHfzPDPcYABABGgJ0bQ&gclid=Cj0KCQjw4bipBhCyARIsAFsie
181. Feynman RP. Surely You’re Joking, Mr. Feynman!. New York: WW
CzU7epQQG4ouyspwn7TPYa7IWh2W0OqZJVXr0qbGhLzshw
Norton, 1985.
182. Hariri AR, Brown SM, Williamson DE et al. Preference for imme OnCk1YEkaAvzLEALw_wcB&sig=AOD64_3vxA5GilCwgEGMqI
diate over delayed rewards is associated with magnitude of 4OhPZk0EBbbQ&q&adurl&ved=2ahUKEwja_52Akf2BAxVWsFY
ventral striatal activity. J Neurosci 2006;26:13213–7. BHcL6CYcQ0Qx6BAgJEAE (date last accessed 20 August 2023).
183. Hansen JE. Scientific reticence and sea level rise. Environ Res 202. Keith DW, Holmes G, Angelo D et al. A process for capturing
Lett 2007;2:024002. er1246875 CO2 from the atmosphere. Joule 2018;2:1573–94.
184. Hansen JE. A slippery slope: how much global warming consti 203. Hansen J, Kharecha P. Cost of carbon capture: can young peo
tutes “dangerous anthropogenic interference?”. Clim Change ple bear the burden? Joule 2018;2:1405–7.
2005;68:269–79. 204. Hardin G. The tragedy of the commons. Science 1968;162:1243–8.
185. Braithewaite RJ. Cover photo for Science 2002;297(5579). 205. Prins G, Rayner S. Time to ditch Kyoto. Nature 2007;449:973–5.
Reprinted in Hansen, J. Defusing the global warming time 206. Creutzig F, Erb KH, Haberl H et al. Considering sustainability
bomb. Sci Amer 2004;290:68–77. thresholds for BECCS in IPCC and biodiversity assessments.
186. Rignot E, Jacobs S, Mouginot J et al. Ice shelf melting around GCB Bioemergy 2021;13:510–5.
Antarctica. Science 2013;341:266–70. 207. Economists’ Statement on Carbon Dividends. https://www.his
187. Rye CD, Naveira Garabato AC, Holland PR et al. Rapid sea-level toryismade.org/ (date last accessed 28 November 2022).
rise along the Antarctic margins in response to increased gla 208. Hansen J. Can Young People Save Democracy and the Planet?
cial discharge. Nature Geosci 2014;7:732–5. https://www.columbia.edu/~jeh1/mailings/2021/20211008_
188. Hansen J, Sato M, Hearty P et al. Ice Melt, Sea Level Rise and YoungPeople.2021.pdf (date last accessed 28 November 2022)
Superstorms: Evidence from Paleoclimate Data, Climate 209. Hayes RB. Nuclear energy myths versus facts support it’s ex
Modeling, and Modern Observations that 2 C Global Warming panded use – a review. Cleaner Ener Sys 2022;2:100009.
is Highly Dangerous. Atmospheric Chemistry and Physics 210. Cao J, Cohen A, Hansen J et al. China-U.S. cooperation to ad
Discussions. 2015. vance nuclear power. Science 2016;353:547–8.
189. Bakker P, Schmittner A, Lenaerts JTM et al. Fate of the Atlantic 211. Ying F. Cooperative competition is possible between China and
Meridional Overturning Circulation: strong decline under con
the U.S. New York Times, 24 November.
tinued warming and Greenland melting. Geophy Res Lett 2016;
212. National Academies of Sciences, Engineering, and Medicine.
43:12252–60.
Reflecting Sunlight: Recommendations for Solar Geoengineering
190. Irvali N, Ninnemann US, Galaasen EV et al. Rapid switches in
Research and Research Governance. https://doi.org/10.17226/
subpolar hydrography and climate during the Last Interglacial
25762 (date last accessed 4 December 2022).
(MIS 5e). Paleoceanography 2012;27:PA2207.
213. Robock A. Volcanic eruptions and climate. Rev Geophys 2000;
191. Blanchon P, Eisenhauer A, Fietzke J et al. Rapid sea-level rise
38:191–219.
and reef back-stepping at the close of the last interglacial high
214. Hansen J, Sato M, Ruedy R et al. A Pinatubo climate Modeling
stand. Nature 2009;458:881–4.
investigation. NATO ASI 1996;I42:234–72.
192. Rye C, Marshall J, Kelley M et al. Antarctic glacial melt as a
215. Tollefson J. Can artificially altered clouds save the Great
driver of recent Southern Ocean climate trends. Geophys Res
Lett 2020;47:e2019GL086892. Barrier Reef? Nature 2021;596:476–8.
193. Silvano A, Rintoul SR, Pena-Molino B et al. Freshening by glacial 216. Latham J, Rasch P, Chen CC et al. Global temperature stabiliza
meltwater enhances melting of ice shelves and reduces forma tion via controlled albedo enhancement of low-level maritime
tion of Antarctic bottom water. Sci Adv 2018;4:eaap9467. clouds. Philos Trans A Math Phys Eng Sci 2008;366:3969–87.
194. Gunn KL, Rintoul SR, England MH et al. Recent reduced abyssal 217. Patrick SM. Reflecting sunlight to reduce climate risk: priorities
overturning and ventilation in the Australian Antarctic Basin. for research and international cooperation. Council on Foreign
Nat Clim Chang 2023;13:537–44. Relations. Special Report No. 93, April 2022 (date last accessed 4
195. Ditlevsen P, Ditlevsen S. Warning of a forthcoming collapse of December 2022).
the Atlantic meridional overturning circulation. Nature Comm 218. Lovins AB. Reinventing Fire. White River Junction, Vermont:
2023;14:4254. Chelsea Green Publishing, 334 p.

Climate Change - Oxford 2023

Uploaded by

Copyright:

Available Formats

Climate Change - Oxford 2023

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Climate Change - Oxford 2023

Uploaded by

Copyright:

Available Formats

Oxford Open Climate Change, 2023, 3(1), kgad008

Global warming in the pipeline

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

5 300 700 300

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

GISS (2014) GISS (2014)

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Cenozoic TS Resulting EECO (Early Eocene Climatic Optimum) TS is �27� C

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Zhu (2019) Zhu (2019)

Cenozoic CO2 " #

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

(a) Holocene CO2 = 260 ppm (b) Holocene CO 2 = 278 ppm

500 450 ppm 500 450 ppm

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

1 net negative Northern Hemisphere forcing and positive Southern

-4 Industrial era aerosols

more widely accepted. Ruddiman’s hypothesis, right or wrong, is

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

(a) GISS (2014) Model (b) GISS (2020) Model

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Global Absorbed Solar Radiation Anomaly (W/m2)

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Best Linear Fit (1970−2010)

Figure 24. Global temperature relative to 1880–1920. Edges of the

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

(a) Global Energy Consumption (b) Global CO 2 Emissions

Global Emissions (GtC/Year)

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

(a) Emissions by Region

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

.02 RCP6.0 RCP4.5

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

Downloaded from https://academic.oup.com/oocc/article/3/1/kgad008/7335889 by guest on 21 November 2023

You might also like