1029
12
This chapter should be cited as:
Collins, M., R. Knutti, J. Arblaster, J.-L. Dufresne, T. Fichefet, P. Friedlingstein, X. Gao, W.J. Gutowski, T. Johns, G.
Krinner, M. Shongwe, C. Tebaldi, A.J. Weaver and M. Wehner, 2013: Long-term Climate Change: Projections, Com-
mitments and Irreversibility. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group
I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K.
Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University
Press, Cambridge, United Kingdom and New York, NY, USA.
Coordinating Lead Authors:
Matthew Collins (UK), Reto Knutti (Switzerland)
Lead Authors:
Julie Arblaster (Australia), Jean-Louis Dufresne (France), Thierry Fichefet (Belgium), Pierre
Friedlingstein (UK/Belgium), Xuejie Gao (China), William J. Gutowski Jr. (USA), Tim Johns (UK),
Gerhard Krinner (France/Germany), Mxolisi Shongwe (South Africa), Claudia Tebaldi (USA),
Andrew J. Weaver (Canada), Michael Wehner (USA)
Contributing Authors:
Myles R. Allen (UK), Tim Andrews (UK), Urs Beyerle (Switzerland), Cecilia M. Bitz (USA),
Sandrine Bony (France), Ben B.B. Booth (UK), Harold E. Brooks (USA), Victor Brovkin (Germany),
Oliver Browne (UK), Claire Brutel-Vuilmet (France), Mark Cane (USA), Robin Chadwick (UK),
Ed Cook (USA), Kerry H. Cook (USA), Michael Eby (Canada), John Fasullo (USA), Erich M.
Fischer (Switzerland), Chris E. Forest (USA), Piers Forster (UK), Peter Good (UK), Hugues Goosse
(Belgium), Jonathan M. Gregory (UK), Gabriele C. Hegerl (UK/Germany), Paul J. Hezel (Belgium/
USA), Kevin I. Hodges (UK), Marika M. Holland (USA), Markus Huber (Switzerland), Philippe
Huybrechts (Belgium), Manoj Joshi (UK), Viatcheslav Kharin (Canada), Yochanan Kushnir (USA),
David M. Lawrence (USA), Robert W. Lee (UK), Spencer Liddicoat (UK), Christopher Lucas
(Australia), Wolfgang Lucht (Germany), Jochem Marotzke (Germany), François Massonnet
(Belgium), H. Damon Matthews (Canada), Malte Meinshausen (Germany), Colin Morice
(UK), Alexander Otto (UK/Germany), Christina M. Patricola (USA), Gwenaëlle Philippon-
Berthier (France), Prabhat (USA), Stefan Rahmstorf (Germany), William J. Riley (USA), Joeri
Rogelj (Switzerland/Belgium), Oleg Saenko (Canada), Richard Seager (USA), Jan Sedláček
(Switzerland), Len C. Shaffrey (UK), Drew Shindell (USA), Jana Sillmann (Canada), Andrew
Slater (USA/Australia), Bjorn Stevens (Germany/USA), Peter A. Stott (UK), Robert Webb (USA),
Giuseppe Zappa (UK/Italy), Kirsten Zickfeld (Canada/Germany)
Review Editors:
Sylvie Joussaume (France), Abdalah Mokssit (Morocco), Karl Taylor (USA), Simon Tett (UK)
Long-term Climate Change:
Projections, Commitments
and Irreversibility
1030
12
Table of Contents
Executive Summary ................................................................... 1031
12.1 Introduction .................................................................... 1034
12.2 Climate Model Ensembles and Sources of
Uncertainty from Emissions to Projections ........... 1035
12.2.1 The Coupled Model Intercomparison Project
Phase 5 and Other Tools .......................................... 1035
12.2.2 General Concepts: Sources of Uncertainties ............ 1035
12.2.3 From Ensembles to Uncertainty Quantification ....... 1040
Box 12.1: Methods to Quantify Model
Agreement in Maps ................................................................. 1041
12.2.4 Joint Projections of Multiple Variables .................... 1044
12.3 Projected Changes in Forcing Agents, Including
Emissions and Concentrations .................................. 1044
12.3.1 Description of Scenarios .......................................... 1045
12.3.2 Implementation of Forcings in Coupled Model
Intercomparison Project Phase 5 Experiments ....... 1047
12.3.3 Synthesis of Projected Global Mean Radiative
Forcing for the 21st Century .................................... 1052
12.4 Projected Climate Change over the
21st Century ................................................................... 1054
12.4.1 Time-Evolving Global Quantities ............................. 1054
12.4.2 Pattern Scaling ........................................................ 1058
12.4.3 Changes in Temperature and Energy Budget ........... 1062
12.4.4 Changes in Atmospheric Circulation ....................... 1071
12.4.5 Changes in the Water Cycle .................................... 1074
12.4.6 Changes in Cryosphere ........................................... 1087
12.4.7 Changes in the Ocean ............................................. 1093
12.4.8 Changes Associated with Carbon Cycle
Feedbacks and Vegetation Cover ............................ 1096
12.4.9 Consistency and Main Differences Between Coupled
Model Intercomparison Project Phase 3/Coupled
Model Intercomparison Project Phase 5 and Special
Report on Emission Scenarios/Representative
Concentration Pathways ........................................ 1099
12.5 Climate Change Beyond 2100, Commitment,
Stabilization and Irreversibility ................................ 1102
12.5.1 Representative Concentration Pathway
Extensions ............................................................... 1102
12.5.2 Climate Change Commitment ................................. 1102
12.5.3 Forcing and Response, Time Scales of Feedbacks .... 1105
12.5.4 Climate Stabilization and Long-term
Climate Targets ....................................................... 1107
Box 12.2: Equilibrium Climate Sensitivity and
Transient Climate Response ................................................... 1110
12.5.5 Potentially Abrupt or Irreversible Changes .............. 1114
References ................................................................................ 1120
Frequently Asked Questions
FAQ 12.1 Why Are So Many Models and Scenarios Used
to Project Climate Change? ................................ 1036
FAQ 12.2 How Will the Earth’s Water Cycle Change? ....... 1084
FAQ 12.3 What Would Happen to Future Climate if We
Stopped Emissions Today? .................................. 1106
1031
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
1
In this Report, the following terms have been used to indicate the assessed likelihood of an outcome or a result: Virtually certain 99–100% probability, Very likely 90–100%,
Likely 66–100%, About as likely as not 33–66%, Unlikely 0–33%, Very unlikely 0–10%, Exceptionally unlikely 0–1%. Additional terms (Extremely likely: 95–100%, More likely
than not >50–100%, and Extremely unlikely 0–5%) may also be used when appropriate. Assessed likelihood is typeset in italics, e.g., very likely (see Section 1.4 and Box TS.1
for more details).
2
In this Report, the following summary terms are used to describe the available evidence: limited, medium, or robust; and for the degree of agreement: low, medium, or high.
A level of confidence is expressed using five qualifiers: very low, low, medium, high, and very high, and typeset in italics, e.g., medium confidence. For a given evidence and
agreement statement, different confidence levels can be assigned, but increasing levels of evidence and degrees of agreement are correlated with increasing confidence (see
Section 1.4 and Box TS.1 for more details).
Executive Summary
This chapter assesses long-term projections of climate change for the
end of the 21st century and beyond, where the forced signal depends
on the scenario and is typically larger than the internal variability of
the climate system. Changes are expressed with respect to a baseline
period of 1986–2005, unless otherwise stated.
Scenarios, Ensembles and Uncertainties
The Coupled Model Intercomparison Project Phase 5 (CMIP5)
presents an unprecedented level of information on which to
base projections including new Earth System Models with a
more complete representation of forcings, new Representative
Concentration Pathways (RCP) scenarios and more output avail-
able for analysis. The four RCP scenarios used in CMIP5 lead to a
total radiative forcing (RF) at 2100 that spans a wider range than that
estimated for the three Special Report on Emission Scenarios (SRES)
scenarios (B1, A1B, A2) used in the Fourth Assessment Report (AR4),
RCP2.6 being almost 2 W m
–2
lower than SRES B1 by 2100. The mag-
nitude of future aerosol forcing decreases more rapidly in RCP sce-
narios, reaching lower values than in SRES scenarios through the 21st
century. Carbon dioxide (CO
2
) represents about 80 to 90% of the total
anthropogenic forcing in all RCP scenarios through the 21st century.
The ensemble mean total effective RFs at 2100 for CMIP5 concen-
tration-driven projections are 2.2, 3.8, 4.8 and 7.6 W m
–2
for RCP2.6,
RCP4.5, RCP6.0 and RCP8.5 respectively, relative to about 1850, and
are close to corresponding Integrated Assessment Model (IAM)-based
estimates (2.4, 4.0, 5.2 and 8.0 W m
–2
). {12.2.1, 12.3, Table 12.1, Fig-
ures 12.1, 12.2, 12.3, 12.4}
New experiments and studies have continued to work towards
a more complete and rigorous characterization of the uncertain-
ties in long-term projections, but the magnitude of the uncer-
tainties has not changed significantly since AR4. There is overall
consistency between the projections based on CMIP3 and CMIP5, for
both large-scale patterns and magnitudes of change. Differences in
global temperature projections are largely attributable to a change in
scenarios. Model agreement and confidence in projections depend on
the variable and spatial and temporal averaging. The well-established
stability of large-scale geographical patterns of change during a tran-
sient experiment remains valid in the CMIP5 models, thus justifying
pattern scaling to approximate changes across time and scenarios
under such experiments. Limitations remain when pattern scaling is
applied to strong mitigation scenarios, to scenarios where localized
forcing (e.g., aerosols) are significant and vary in time and for varia-
bles other than average temperature and precipitation. {12.2.2, 12.2.3,
12.4.2, 12.4.9, Figures 12.10, 12.39, 12.40, 12.41}
Projections of Temperature Change
Global mean temperatures will continue to rise over the 21st
century if greenhouse gas (GHG) emissions continue unabat-
ed. Under the assumptions of the concentration-driven RCPs, global
mean surface temperatures for 2081–2100, relative to 1986–2005 will
likely
1
be in the 5 to 95% range of the CMIP5 models; 0.3°C to 1.7°C
(RCP2.6), 1.1°C to 2.6°C (RCP4.5), 1.4°C to 3.1°C (RCP6.0), 2.6°C to
4.8°C (RCP8.5). Global temperatures averaged over the period 2081–
2100 are projected to likely exceed 1.5°C above 1850-1900 for RCP4.5,
RCP6.0 and RCP8.5 (high confidence), are likely to exceed 2°C above
1850-1900 for RCP6.0 and RCP8.5 (high confidence) and are more
likely than not to exceed 2°C for RCP4.5 (medium confidence). Temper-
ature change above 2°C under RCP2.6 is unlikely (medium confidence).
Warming above 4°C by 2081–2100 is unlikely in all RCPs (high confi-
dence) except for RCP8.5, where it is about as likely as not (medium
confidence). {12.4.1, Tables 12.2, 12.3, Figures 12.5, 12.8}
Temperature change will not be regionally uniform. There is very
high confidence
2
that globally averaged changes over land will exceed
changes over the ocean at the end of the 21st century by a factor that
is likely in the range 1.4 to 1.7. In the absence of a strong reduction
in the Atlantic Meridional Overturning, the Arctic region is project-
ed to warm most (very high confidence). This polar amplification is
not found in Antarctic regions due to deep ocean mixing, ocean heat
uptake and the persistence of the Antarctic ice sheet. Projected region-
al surface air temperature increase has minima in the North Atlantic
and Southern Oceans in all scenarios. One model exhibits marked cool-
ing in 2081–2100 over large parts of the Northern Hemisphere (NH),
and a few models indicate slight cooling locally in the North Atlantic.
Atmospheric zonal mean temperatures show warming throughout the
troposphere, especially in the upper troposphere and northern high
latitudes, and cooling in the stratosphere. {12.4.2, 12.4.3, Table 12.2,
Figures 12.9, 12.10, 12.11, 12.12}
It is virtually certain that, in most places, there will be more hot
and fewer cold temperature extremes as global mean temper-
atures increase. These changes are expected for events defined as
extremes on both daily and seasonal time scales. Increases in the fre-
quency, duration and magnitude of hot extremes along with heat stress
are expected; however, occasional cold winter extremes will continue to
occur. Twenty-year return values of low temperature events are project-
ed to increase at a rate greater than winter mean temperatures in most
regions, with the largest changes in the return values of low tempera-
tures at high latitudes. Twenty-year return values for high temperature
events are projected to increase at a rate similar to or greater than the
rate of increase of summer mean temperatures in most regions. Under
RCP8.5 it is likely that, in most land regions, a current 20-year high
temperature event will occur more frequently by the end of the 21st
1032
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
century (at least doubling its frequency, but in many regions becoming
an annual or 2-year event) and a current 20-year low temperature event
will become exceedingly rare. {12.4.3, Figures 12.13, 12.14}
Changes in Atmospheric Circulation
Mean sea level pressure is projected to decrease in high lati-
tudes and increase in the mid-latitudes as global temperatures
rise. In the tropics, the Hadley and Walker Circulations are likely
to slow down. Poleward shifts in the mid-latitude jets of about 1
to 2 degrees latitude are likely at the end of the 21st century under
RCP8.5 in both hemispheres (medium confidence), with weaker shifts
in the NH. In austral summer, the additional influence of stratospheric
ozone recovery in the Southern Hemisphere opposes changes due to
GHGs there, though the net response varies strongly across models and
scenarios. Substantial uncertainty and thus low confidence remains in
projecting changes in NH storm tracks, especially for the North Atlantic
basin. The Hadley Cell is likely to widen, which translates to broad-
er tropical regions and a poleward encroachment of subtropical dry
zones. In the stratosphere, the Brewer–Dobson circulation is likely to
strengthen. {12.4.4, Figures 12.18, 12.19, 12.20}
Changes in the Water Cycle
It is virtually certain that, in the long term, global precipitation
will increase with increased global mean surface temperature.
Global mean precipitation will increase at a rate per degree Celsius
smaller than that of atmospheric water vapour. It will likely increase by
1 to 3% °C
–1
for scenarios other than RCP2.6. For RCP2.6 the range of
sensitivities in the CMIP5 models is 0.5 to 4% °C
–1
at the end of the
21st century. {12.4.1, Figures 12.6, 12.7}
Changes in average precipitation in a warmer world will exhibit
substantial spatial variation. Some regions will experience
increases, other regions will experience decreases and yet
others will not experience significant changes at all. There is
high confidence that the contrast of annual mean precipitation
between dry and wet regions and that the contrast between
wet and dry seasons will increase over most of the globe as
temperatures increase. The general pattern of change indicates that
high latitude land masses are likely to experience greater amounts
of precipitation due to the increased specific humidity of the warmer
troposphere as well as increased transport of water vapour from the
tropics by the end of this century under the RCP8.5 scenario. Many
mid-latitude and subtropical arid and semi-arid regions will likely
experience less precipitation and many moist mid-latitude regions will
likely experience more precipitation by the end of this century under
the RCP8.5 scenario. Globally, for short-duration precipitation events, a
shift to more intense individual storms and fewer weak storms is likely
as temperatures increase. Over most of the mid-latitude land-masses
and over wet tropical regions, extreme precipitation events will very
likely be more intense and more frequent in a warmer world. The global
average sensitivity of the 20-year return value of the annual maximum
daily precipitation increases ranges from 4% °C
–1
of local temperature
increase (average of CMIP3 models) to 5.3%
o
C
–1
of local tempera-
ture increase (average of CMIP5 models) but regionally there are wide
variations. {12.4.5, Figures 12.10, 12.22, 12.26, 12.27}
Annual surface evaporation is projected to increase as global
temperatures rise over most of the ocean and is projected to
change over land following a similar pattern as precipitation.
Decreases in annual runoff are likely in parts of southern Europe, the
Middle East, and southern Africa by the end of the 21st century under
the RCP8.5 scenario. Increases in annual runoff are likely in the high
northern latitudes corresponding to large increases in winter and
spring precipitation by the end of the 21st century under the RCP8.5
scenario. Regional to global-scale projected decreases in soil moisture
and increased risk of agricultural drought are likely in presently dry
regions and are projected with medium confidence by the end of the
21st century under the RCP8.5 scenario. Prominent areas of projected
decreases in evaporation include southern Africa and north western
Africa along the Mediterranean. Soil moisture drying in the Mediterra-
nean, southwest USA and southern African regions is consistent with
projected changes in Hadley Circulation and increased surface tem-
peratures, so surface drying in these regions as global temperatures
increase is likely with high confidence by the end of this century under
the RCP8.5 scenario. In regions where surface moistening is projected,
changes are generally smaller than natural variability on the 20-year
time scale. {12.4.5, Figures 12.23, 12.24, 12.25}
Changes in Cryosphere
It is very likely that the Arctic sea ice cover will continue shrink-
ing and thinning year-round in the course of the 21st century as
global mean surface temperature rises. At the same time, in the
Antarctic, a decrease in sea ice extent and volume is expected,
but with low confidence. Based on the CMIP5 multi-model ensem-
ble, projections of average reductions in Arctic sea ice extent for 2081–
2100 compared to 1986–2005 range from 8% for RCP2.6 to 34% for
RCP8.5 in February and from 43% for RCP2.6 to 94% for RCP8.5 in
September (medium confidence). A nearly ice-free Arctic Ocean (sea ice
extent less than 1 × 10
6
km
2
for at least 5 consecutive years) in Septem-
ber before mid-century is likely under RCP8.5 (medium confidence),
based on an assessment of a subset of models that most closely repro-
duce the climatological mean state and 1979–2012 trend of the Arctic
sea ice cover. Some climate projections exhibit 5- to 10-year periods of
sharp summer Arctic sea ice decline—even steeper than observed over
the last decade—and it is likely that such instances of rapid ice loss
will occur in the future. There is little evidence in global climate models
of a tipping point (or critical threshold) in the transition from a peren-
nially ice-covered to a seasonally ice-free Arctic Ocean beyond which
further sea ice loss is unstoppable and irreversible. In the Antarctic, the
CMIP5 multi-model mean projects a decrease in sea ice extent that
ranges from 16% for RCP2.6 to 67% for RCP8.5 in February and from
8% for RCP2.6 to 30% for RCP8.5 in September for 2081–2100 com-
pared to 1986–2005. There is, however, low confidence in those values
as projections because of the wide inter-model spread and the inability
of almost all of the available models to reproduce the mean annual
cycle, interannual variability and overall increase of the Antarctic sea
ice areal coverage observed during the satellite era. {12.4.6, 12.5.5,
Figures 12.28, 12.29, 12.30, 12.31}
It is very likely that NH snow cover will reduce as global tem-
peratures rise over the coming century. A retreat of permafrost
extent with rising global temperatures is virtually certain. Snow
1033
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
cover changes result from precipitation and ablation changes, which
are sometimes opposite. Projections of the NH spring snow covered
area by the end of the 21st century vary between a decrease of 7%
(RCP2.6) and a decrease of 25% (RCP8.5), with a pattern that is fairly
consistent between models. The projected changes in permafrost are a
response not only to warming but also to changes in snow cover, which
exerts a control on the underlying soil. By the end of the 21st cen-
tury, diagnosed near-surface permafrost area is projected to decrease
by between 37% (RCP2.6) and 81% (RCP8.5) (medium confidence).
{12.4.6, Figures 12.32, 12.33}
Changes in the Ocean
The global ocean will warm in all RCP scenarios. The strongest
ocean warming is projected for the surface in subtropical and tropi-
cal regions. At greater depth the warming is projected to be most
pronounced in the Southern Ocean. Best estimates of ocean warm-
ing in the top onehundred meters are about 0.6°C (RCP2.6) to 2.0°C
(RCP8.5), and about 0.3°C (RCP2.6) to 0.6°C(RCP8.5) at a depth of
about 1 km by the end of the 21st century. For RCP4.5 by the end of the
21st century, half of the energy taken up by the ocean is in the upper-
most 700 m and 85% is in the uppermost 2000 m. Due to the long time
scales of this heat transfer from the surface to depth, ocean warming
will continue for centuries, even if GHG emissions are decreased or
concentrations kept constant. {12.4.7, 12.5.2–12.5.4, Figure 12.12}
It is very likely that the Atlantic Meridional Overturning Circu-
lation (AMOC) will weaken over the 21st century but it is very
unlikely that the AMOC will undergo an abrupt transition or col-
lapse in the 21st century. Best estimates and ranges for the reduc-
tion from CMIP5 are 11% (1 to 24%) in RCP2.6 and 34% (12 to 54%)
in RCP8.5. There is low confidence in assessing the evolution of the
AMOC beyond the 21st century. {12.4.7, Figure 12.35}
Carbon Cycle
When forced with RCP8.5 CO
2
emissions, as opposed to the
RCP8.5 CO
2
concentrations, 11 CMIP5 Earth System Models with
interactive carbon cycle simulate, on average, a 50 ppm (min to
max range –140 to +210 ppm) larger atmospheric CO
2
concen-
tration and 0.2°C (min to max range –0.4 to +0.9°C) larger global
surface temperature increase by 2100. {12.4.8, Figures 12.36, 12.37}
Long-term Climate Change, Commitment and Irreversibility
Global temperature equilibrium would be reached only after
centuries to millennia if RF were stabilized. Continuing GHG emis-
sions beyond 2100, as in the RCP8.5 extension, induces a total RF above
12 W m
–2
by 2300. Sustained negative emissions beyond 2100, as in
RCP2.6, induce a total RF below 2 W m
–2
by 2300. The projected warm-
ing for 2281–2300, relative to 1986–2005, is 0.0°C to 1.2°C for RCP2.6
and 3.0°C to 12.6°C for RCP8.5 (medium confidence). In much the same
way as the warming to a rapid increase of forcing is delayed, the cooling
after a decrease of RF is also delayed. {12.5.1, Figures 12.43, 12.44}
A large fraction of climate change is largely irreversible on
human time scales, unless net anthropogenic CO
2
emissions
were strongly negative over a sustained period. For scenarios
driven by CO
2
alone, global average temperature is projected to
remain approximately constant for many centuries following a com-
plete cessation of emissions. The positive commitment from CO
2
may
be enhanced by the effect of an abrupt cessation of aerosol emissions,
which will cause warming. By contrast, cessation of emission of short-
lived GHGs will contribute a cooling. {12.5.3, 12.5.4, Figures 12.44,
12.45, 12.46, FAQ 12.3}
Equilibrium Climate Sensitivity and Transient Climate
Response
Estimates of the equilibrium climate sensitivity (ECS) based on
observed climate change, climate models and feedback analy-
sis, as well as paleoclimate evidence indicate that ECS is likely
in the range 1.5°C to 4.5°C with high confidence, extreme-
ly unlikely less than 1°C (high confidence) and very unlikely
greater than 6°C (medium confidence). The transient climate
response (TCR) is likely in the range 1°C to 2.5ºC and extremely
unlikely greater than 3°C, based on observed climate change
and climate models. {Box 12.2, Figures 1, 2}
Climate Stabilization
The principal driver of long-term warming is total emissions
of CO
2
and the two quantities are approximately linearly
related. The global mean warming per 1000 PgC (transient cli-
mate response to cumulative carbon emissions (TCRE)) is likely
between 0.8°C to 2.5°C per 1000 PgC, for cumulative emissions
less than about 2000 PgC until the time at which temperatures
peak. To limit the warming caused by anthropogenic CO
2
emissions
alone to be likely less than 2°C relative to the period 1861-1880, total
CO
2
emissions from all anthropogenic sources would need to be limit-
ed to a cumulative budget of about 1000 PgC since that period. About
half [445 to 585 PgC] of this budget was already emitted by 2011.
Accounting for projected warming effect of non-CO
2
forcing, a possible
release of GHGs from permafrost or methane hydrates, or requiring
a higher likelihood of temperatures remaining below 2°C, all imply a
lower budget. {12.5.4, Figures 12.45, 12.46, Box 12.2}
Some aspects of climate will continue to change even if temper-
atures are stabilized. Processes related to vegetation change, chang-
es in the ice sheets, deep ocean warming and associated sea level rise
and potential feedbacks linking for example ocean and the ice sheets
have their own intrinsic long time scales and may result in significant
changes hundreds to thousands of years after global temperature is
stabilized. {12.5.2 to 12.5.4}
Abrupt Change
Several components or phenomena in the climate system could
potentially exhibit abrupt or nonlinear changes, and some are
known to have done so in the past. Examples include the AMOC,
Arctic sea ice, the Greenland ice sheet, the Amazon forest and mon-
soonal circulations. For some events, there is information on potential
consequences, but in general there is low confidence and little con-
sensus on the likelihood of such events over the 21st century. {12.5.5,
Table 12.4}
1034
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
12.1 Introduction
Projections of future climate change are not like weather forecasts.
It is not possible to make deterministic, definitive predictions of how
climate will evolve over the next century and beyond as it is with short-
term weather forecasts. It is not even possible to make projections of
the frequency of occurrence of all possible outcomes in the way that it
might be possible with a calibrated probabilistic medium-range weath-
er forecast. Projections of climate change are uncertain, first because
they are dependent primarily on scenarios of future anthropogenic
and natural forcings that are uncertain, second because of incomplete
understanding and imprecise models of the climate system and finally
because of the existence of internal climate variability. The term cli-
mate projection tacitly implies these uncertainties and dependencies.
Nevertheless, as greenhouse gas (GHG) concentrations continue to
rise, we expect to see future changes to the climate system that are
greater than those already observed and attributed to human activi-
ties. It is possible to understand future climate change using models
and to use models to characterize outcomes and uncertainties under
specific assumptions about future forcing scenarios.
This chapter assesses climate projections on time scales beyond those
covered in Chapter 11, that is, beyond the mid-21st century. Informa-
tion from a range of different modelling tools is used here; from simple
energy balance models, through Earth System Models of Intermediate
Complexity (EMICs) to complex dynamical climate and Earth System
Models (ESMs). These tools are evaluated in Chapter 9 and, where pos-
sible, the evaluation is used in assessing the validity of the projections.
This chapter also summarizes some of the information on leading-order
measures of the sensitivity of the climate system from other chapters
and discusses the relevance of these measures for climate projections,
commitments and irreversibility.
Since the AR4 (Meehl et al., 2007b) there have been a number of
advances:
New scenarios of future forcings have been developed to replace
the Special Report on Emissions Scenarios (SRES). The Represen-
tative Concentration Pathways (RCPs, see Section 12.3) (Moss et
al., 2010), have been designed to cover a wide range of possible
magnitudes of climate change in models rather than being derived
sequentially from storylines of socioeconomic futures. The aim is
to provide a range of climate responses while individual socioeco-
nomic scenarios may be derived, scaled and interpolated (some
including explicit climate policy). Nevertheless, many studies that
have been performed since AR4 have used SRES and, where appro-
priate, these are assessed. Simplified scenarios of future change,
developed strictly for understanding the response of the climate
system rather than to represent realistic future outcomes, are also
synthesized and the understanding of leading-order measures of
climate response such as the equilibrium climate sensitivity (ECS)
and the transient climate response (TCR) are assessed.
New models have been developed with higher spatial resolution,
with better representation of processes and with the inclusion of
more processes, in particular processes that are important in simu-
lating the carbon cycle of the Earth. In these models, emissions of
GHGs may be specified and these gases may be chemically active
in the atmosphere or be exchanged with pools in terrestrial and
oceanic systems before ending up as an airborne concentration
(see Figure 10.1 of AR4).
New types of model experiments have been performed, many
coordinated by the Coupled Model Intercomparison Project Phase
5 (CMIP5) (Taylor et al., 2012), which exploit the addition of these
new processes. Models may be driven by emissions of GHGs, or by
their concentrations with different Earth System feedback loops
cut. This allows the separate assessment of different feedbacks in
the system and of projections of physical climate variables and
future emissions.
Techniques to assess and quantify uncertainties in projections
have been further developed but a full probabilistic quantifica-
tion remains difficult to propose for most quantities, the exception
being global, temperature-related measures of the system sensitiv-
ity to forcings, such as ECS and TCR. In those few cases, projections
are presented in the form of probability density functions (PDFs).
We make the distinction between the spread of a multi-model
ensemble, an ad hoc measure of the possible range of projections
and the quantification of uncertainty that combines information
from models and observations using statistical algorithms. Just like
climate models, different techniques for quantifying uncertainty
exist and produce different outcomes. Where possible, different
estimates of uncertainty are compared.
Although not an advance, as time has moved on, the baseline period
from which climate change is expressed has also moved on (a common
baseline period of 1986–2005 is used throughout, consistent with
the 2006 start-point for the RCP scenarios). Hence climate change is
expressed as a change with respect to a recent period of history, rather
than a time before significant anthropogenic influence. It should be
borne in mind that some anthropogenically forced climate change had
already occurred by the 1986–2005 period (see Chapter 10).
The focus of this chapter is on global and continental/ocean basin-scale
features of climate. For many aspects of future climate change, it is
possible to discuss generic features of projections and the processes
that underpin them for such large scales. Where interesting or unique
changes have been investigated at smaller scales, and there is a level
of agreement between different studies of those smaller-scale changes,
these may also be assessed in this chapter, although where changes are
linked to climate phenomena such as El Niño, readers are referred to
Chapter 14. Projections of atmospheric composition, chemistry and air
quality for the 21st century are assessed in Chapter 11, except for CO
2
which is assessed in this chapter. An innovation for AR5 is Annex I: Atlas
of Global and Regional Climate Projections, a collection of global and
regional maps of projected climate changes derived from model output.
A detailed commentary on each of the maps presented in Annex I is not
provided here, but some discussion of generic features is provided.
Projections from regional models driven by boundary conditions from
global models are not extensively assessed but may be mentioned
in this chapter. More detailed regional information may be found in
Chapter 14 and is also now assessed in the Working Group II report,
where it can more easily be linked to impacts.
1035
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
12.2 Climate Model Ensembles and Sources of
Uncertainty from Emissions to Projections
12.2.1 The Coupled Model Intercomparison Project
Phase 5 and Other Tools
Many of the figures presented in this chapter and in others draw
on data collected as part of CMIP5 (Taylor et al., 2012). The project
involves the worldwide coordination of ESM experiments including the
coordination of input forcing fields, diagnostic output and the host-
ing of data in a distributed archive. CMIP5 has been unprecedented
in terms of the number of modelling groups and models participating,
the number of experiments performed and the number of diagnostics
collected. The archive of model simulations began being populated by
mid-2011 and continued to grow during the writing of AR5. The pro-
duction of figures for this chapter draws on a fixed database of simu-
lations and variables that was available on 15 March 2013 (the same
as the cut-off date for the acceptance of the publication of papers).
Different figures may use different subsets of models and there are
unequal numbers of models that have produced output for the differ-
ent RCP scenarios. Figure 12.1 gives a summary of which output was
available from which model for which scenario. Where multiple runs
Model/Variable taspsl pr clthurshussevspsbl rsut rlut rtmt rsdt mrro mrso tsltauamsft.yz sossic snctas_day pr_day
ACCESS1-0
ACCESS1-3
bcc-csm1-1
bcc-csm1-1-m
BNU-ESM
CanESM2
CCSM4
CESM1-BGC
CESM1-CAM5
CESM1-WACCM
CMCC-CESM
CMCC-CM
CMCC-CMS
CNRM-CM5
CSIRO-Mk3-6-0
EC-EARTH
FGOALS-g2
FIO-ESM
GFDL-CM3
GFDL-ESM2G
GFDL-ESM2M
GISS-E2-H-CC
GISS-E2-H-P1
GISS-E2-H-P2
GISS-E2-H-P3
GISS-E2-R-CC
GISS-E2-R-P1
GISS-E2-R-P2
GISS-E2-R-P3
HadGEM2-AO
HadGEM2-CC
HadGEM2-ES
inmcm4
IPSL-CM5A-LR
IPSL-CM5A-MR
IPSL-CM5B-LR
MIROC5
MIROC-ESM
MIROC-ESM-CHEM
MPI-ESM-LR
MPI-ESM-MR
MPI-ESM-P
MRI-CGCM3
NorESM1-M
NorESM1-ME
0 ensemble
1 ensemble
2 ensembles
3 ensembles
4 ensembles
5 or more ensembles
Figure 12.1 | A summary of the output used to make the CMIP5 figures in this chapter (and some figures in Chapter 11). The climate variable names run along the horizontal axis
and use the standard abbreviations in the CMIP5 protocol (Taylor et al., 2012, and online references therein). The climate model names run along the vertical axis. In each box the
shading indicates the number of ensemble members available for historical, RCP2.6, RCP4.5, RCP6.0, RCP8.5 and pre-industrial control experiments, although only one ensemble
member per model is used in the relevant figures.
are performed with exactly the same model but with different initial
conditions, we choose only one ensemble member (usually the first but
in cases where that was not available, the first available member is
chosen) in order not to weight models with more ensemble members
than others unduly in the multi-model synthesis. Rather than give an
exhaustive account of which models were used to make which figures,
this summary information is presented as a guide to readers.
In addition to output from CMIP5, information from a coordinated
set of simulations with EMICs is also used (Zickfeld et al., 2013) to
investigate long-term climate change beyond 2100. Even more sim-
plified energy balance models or emulation techniques are also used,
mostly to estimate responses where ESM experiments are not availa-
ble (Meinshausen et al., 2011a; Good et al., 2013). An evaluation of
the models used for projections is provided in Chapter 9 of this Report.
12.2.2 General Concepts: Sources of Uncertainties
The understanding of the sources of uncertainty affecting future cli-
mate change projections has not substantially changed since AR4, but
many experiments and studies since then have proceeded to explore
and characterize those uncertainties further. A full characterization,
1036
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Frequently Asked Questions
FAQ 12.1 | Why Are So Many Models and Scenarios Used to Project Climate Change?
Future climate is partly determined by the magnitude of future emissions of greenhouse gases, aerosols and other
natural and man-made forcings. These forcings are external to the climate system, but modify how it behaves.
Future climate is shaped by the Earth’s response to those forcings, along with internal variability inherent in the
climate system. A range of assumptions about the magnitude and pace of future emissions helps scientists develop
different emission scenarios, upon which climate model projections are based. Different climate models, mean-
while, provide alternative representations of the Earth’s response to those forcings, and of natural climate variabil-
ity. Together, ensembles of models, simulating the response to a range of different scenarios, map out a range of
possible futures, and help us understand their uncertainties.
Predicting socioeconomic development is arguably even more difficult than predicting the evolution of a physical
system. It entails predicting human behaviour, policy choices, technological advances, international competition
and cooperation. The common approach is to use scenarios of plausible future socioeconomic development, from
which future emissions of greenhouse gases and other forcing agents are derived. It has not, in general, been pos-
sible to assign likelihoods to individual forcing scenarios. Rather, a set of alternatives is used to span a range of
possibilities. The outcomes from different forcing scenarios provide policymakers with alternatives and a range of
possible futures to consider.
Internal fluctuations in climate are spontaneously generated by interactions between components such as the
atmosphere and the ocean. In the case of near-term climate change, they may eclipse the effect of external per-
turbations, like greenhouse gas increases (see Chapter 11). Over the longer term, however, the effect of external
forcings is expected to dominate instead. Climate model simulations project that, after a few decades, different
scenarios of future anthropogenic greenhouse gases and other forcing agents—and the climate system’s response
to them—will differently affect the change in mean global temperature (FAQ 12.1, Figure 1, left panel). Therefore,
evaluating the consequences of those various scenarios and responses is of paramount importance, especially when
policy decisions are considered.
Climate models are built on the basis of the physical principles governing our climate system, and empirical under-
standing, and represent the complex, interacting processes needed to simulate climate and climate change, both
past and future. Analogues from past observations, or extrapolations from recent trends, are inadequate strategies
for producing projections, because the future will not necessarily be a simple continuation of what we have seen
thus far.
Although it is possible to write down the equations of fluid motion that determine the behaviour of the atmo-
sphere and ocean, it is impossible to solve them without using numerical algorithms through computer model
simulation, similarly to how aircraft engineering relies on numerical simulations of similar types of equations. Also,
many small-scale physical, biological and chemical processes, such as cloud processes, cannot be described by those
equations, either because we lack the computational ability to describe the system at a fine enough resolution
to directly simulate these processes or because we still have a partial scientific understanding of the mechanisms
driving these processes. Those need instead to be approximated by so-called parameterizations within the climate
models, through which a mathematical relation between directly simulated and approximated quantities is estab-
lished, often on the basis of observed behaviour.
There are various alternative and equally plausible numerical representations, solutions and approximations for
modelling the climate system, given the limitations in computing and observations. This diversity is considered a
healthy aspect of the climate modelling community, and results in a range of plausible climate change projections
at global and regional scales. This range provides a basis for quantifying uncertainty in the projections, but because
the number of models is relatively small, and the contribution of model output to public archives is voluntary,
the sampling of possible futures is neither systematic nor comprehensive. Also, some inadequacies persist that are
common to all models; different models have different strength and weaknesses; it is not yet clear which aspects
of the quality of the simulations that can be evaluated through observations should guide our evaluation of future
model simulations. (continued on next page)
1037
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
FAQ 12.1 (continued)
Models of varying complexity are commonly used for different projection problems. A faster model with lower
resolution, or a simplified description of some climate processes, may be used in cases where long multi-century
simulations are required, or where multiple realizations are needed. Simplified models can adequately represent
large-scale average quantities, like global average temperature, but finer details, like regional precipitation, can be
simulated only by complex models.
The coordination of model experiments and output by groups such as the Coupled Model Intercomparison Project
(CMIP), the World Climate Research Program and its Working Group on Climate Models has seen the science com-
munity step up efforts to evaluate the ability of models to simulate past and current climate and to compare future
climate change projections. The ‘multi-model’ approach is now a standard technique used by the climate science
community to assess projections of a specific climate variable.
FAQ 12.1, Figure 1, right panels, shows the temperature response by the end of the 21st century for two illustrative
models and the highest and lowest RCP scenarios. Models agree on large-scale patterns of warming at the surface,
for example, that the land is going to warm faster than ocean, and the Arctic will warm faster than the tropics. But
they differ both in the magnitude of their global response for the same scenario, and in small scale, regional aspects
of their response. The magnitude of Arctic amplification, for instance, varies among different models, and a subset
of models show a weaker warming or slight cooling in the North Atlantic as a result of the reduction in deepwater
formation and shifts in ocean currents.
There are inevitable uncertainties in future external forcings, and the climate system’s response to them, which
are further complicated by internally generated variability. The use of multiple scenarios and models have become
a standard choice in order to assess and characterize them, thus allowing us to describe a wide range of possible
future evolutions of the Earth’s climate.
FAQ 12.1, Figure 1 | Global mean temperature change averaged across all Coupled Model Intercomparison Project Phase 5 (CMIP5) models (relative to 1986–2005)
for the four Representative Concentration Pathway (RCP) scenarios: RCP2.6 (dark blue), RCP4.5 (light blue), RCP6.0 (orange) and RCP8.5 (red); 32, 42, 25 and 39
models were used respectively for these 4 scenarios. Likely ranges for global temperature change by the end of the 21st centuryare indicated by vertical bars. Note that
these ranges apply to the difference between two 20-year means, 2081–2100 relative to 1986–2005, which accounts for the bars being centred at a smaller value than
the end point of the annual trajectories. For the highest (RCP8.5) and lowest (RCP2.6) scenario, illustrative maps of surface temperature change at the end of the 21st
century (2081–2100 relative to 1986–2005) are shown for two CMIP5 models. These models are chosen to show a rather broad range of response, but this particular
set is not representative of any measure of model response uncertainty.
Model mean global
mean temperature
change for high
emission scenario
RCP8.5
Model mean global
mean temperature
change for low
emission scenario
RCP2.6
Global surface temperature change (°C)
Possible temperature responses in 2081-2100 to
high emission scenario RCP8.5
Possible temperature responses in 2081-2100 to
low emission scenario RCP2.6
-2 -1.5 -1-0.5 00.5 11.5 23457911
(°C)
1038
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
qualitative and even more so quantitative, involves much more than a
measure of the range of model outcomes, because additional sources
of information (e.g., observational constraints, model evaluation, expert
judgement) lead us to expect that the uncertainty around the future
climate state does not coincide straightforwardly with those ranges.
In fact, in this chapter we highlight wherever relevant the distinction
between model uncertainty evaluation, which encompasses the under-
standing that models have intrinsic shortcoming in fully and accurately
representing the real system, and cannot all be considered independent
of one another (Knutti et al., 2013), and a simpler descriptive quantifi-
cation, based on the range of outcomes from the ensemble of models.
Uncertainty affecting mid- to long-term projections of climatic changes
stems from distinct but possibly interacting sources. Figure 12.2 shows
a schematic of the chain from scenarios, through ESMs to projections.
Uncertainties affecting near-term projections of which some aspect
are also relevant for longer-term projections are discussed in Section
11.3.1.1 and shown in Figure 11.8.
Future anthropogenic emissions of GHGs, aerosol particles and other
forcing agents such as land use change are dependent on socioec-
onomic factors including global geopolitical agreements to control
those emissions. Systematic studies that attempt to quantify the likely
ranges of anthropogenic emission have been undertaken (Sokolov et
al., 2009) but it is more common to use a scenario approach of dif-
ferent but plausible—in the sense of technically feasible—pathways,
leading to the concept of scenario uncertainty. AR4 made extensive
use of the SRES scenarios (IPCC, 2000) developed using a sequential
approach, that is, socioeconomic factors feed into emissions scenarios
which are then used either to directly force the climate models or to
determine concentrations of GHGs and other agents required to drive
these models. This report also assesses outcomes of simulations that
use the new RCP scenarios, developed using a parallel process (Moss
et al., 2010) whereby different targets in terms of RF at 2100 were
selected (2.6, 4.5, 6.0 and 8.5 W m
–2
) and GHG and aerosol emissions
consistent with those targets, and their corresponding socioeconom-
ic drivers were developed simultaneously (see Section 12.3). Rather
than being identified with one socioeconomic storyline, RCP scenarios
are consistent with many possible economic futures (in fact, different
combinations of GHG and aerosol emissions can lead to the same
RCP). Their development was driven by the need to produce scenari-
os that could be input to climate model simulations more expediently
while corresponding socioeconomic scenarios would be developed in
parallel, and to produce a wide range of model responses that may be
scaled and interpolated to estimate the response under other scenari-
os, involving different measures of adaptation and mitigation.
In terms of the uncertainties related to the RCP emissions scenarios,
the following issues can be identified:
No probabilities or likelihoods have been attached to the alterna-
tive RCP scenarios (as was the case for SRES scenarios). Each of
them should be considered plausible, as no study has questioned
their technical feasibility (see Chapter 1).
Target Radiative
Forcing
Concentrations
Emissions
Diagnosed Radiative
Forcing
Earth System
Models
Diagnosed
Emissions
Climate Projections
Representative
Concentration Pathway (RCP)
Figure 12.2 | Links in the chain from scenarios, through models to climate projections. The Representative Concentration Pathways (RCPs) are designed to sample a range of
radiative forcing (RF) of the climate system at 2100. The RCPs are translated into both concentrations and emissions of greenhouse gases using Integrated Assessment Models
(IAMs). These are then used as inputs to dynamical Earth System Models (ESMs) in simulations that are either concentration-driven (the majority of projection experiments) or
emissions-driven (only for RCP8.5). Aerosols and other forcing factors are implemented in different ways in each ESM. The ESM projections each have a potentially different RF,
which may be viewed as an output of the model and which may not correspond to precisely the level of RF indicated by the RCP nomenclature. Similarly, for concentration-driven
experiments, the emissions consistent with those concentrations diagnosed from the ESM may be different from those specified in the RCP (diagnosed from the IAM). Different
models produce different responses even under the same RF. Uncertainty propagates through the chain and results in a spread of ESM projections. This spread is only one way
of assessing uncertainty in projections. Alternative methods, which combine information from simple and complex models and observations through statistical models or expert
judgement, are also used to quantify that uncertainty.
1039
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Despite the naming of the RCPs in terms of their target RF at 2100
or at stabilization (Box 1.1), climate models translate concentra-
tions of forcing agents into RF in different ways due to their differ-
ent structural modelling assumptions. Hence a model simulation
of RCP6.0 may not attain exactly a RF of 6 W m
–2
; more accurately,
an RCP6.0 forced model experiment may not attain exactly the
same RF as was intended by the specification of the RCP6.0 forc-
ing inputs. Thus in addition to the scenario uncertainty there is
RF uncertainty in the way the RCP scenarios are implemented in
climate models.
Some model simulations are concentration-driven (GHG concen-
trations are specified) whereas some models, which have Earth
Systems components, convert emission scenarios into concen-
trations and are termed emissions-driven. Different ESMs driven
by emissions may produce different concentrations of GHGs and
aerosols because of differences in the representation and/or
parameterization of the processes responsible for the conversion
of emissions into concentrations. This aspect may be considered a
facet of forcing uncertainty, or may be compounded in the category
of model uncertainty, which we discuss below. Also, aerosol load-
ing and land use changes are not dictated intrinsically by the RCP
specification. Rather, they are a result of the Integrated Assessment
Model that created the emission pathway for a given RCP.
SRES and RCPs account for future changes only in anthropogenic forc-
ings. With regard to solar forcing, the 1985–2005 solar cycle is repeat-
ed. Neither projections of future deviations from this solar cycle, nor
future volcanic RF and their uncertainties are considered.
Any climate projection is subject to sampling uncertainties that arise
because of internal variability. In this chapter, the prediction of, for
example, the amplitude or phase of some mode of variability that may
be important on long time scales is not addressed (see Sections 11.2
and 11.3). Any climate variable projection derived from a single simu-
lation of an individual climate model will be affected by internal varia-
bility (stemming from the chaotic nature of the system), whether it be
a variable that involves a long time average (e.g., 20 years), a snapshot
in time or some more complex diagnostic such as the variance comput-
ed from a time series over many years. No amount of time averaging
can reduce internal variability to zero, although for some EMICs and
simplified models, which may be used to reproduce the results of more
complex model simulations, the representation of internal variability
is excluded from the model specification by design. For different
variables, and different spatial and time scale averages, the relative
importance of internal variability in comparison with other sources of
uncertainty will be different. In general, internal variability becomes
more important on shorter time scales and for smaller scale variables
(see Section 11.3 and Figure 11.2). The concept of signal-to-noise ratio
may be used to quantify the relative magnitude of the forced response
(signal) versus internal variability (noise). Internal variability may be
sampled and estimated explicitly by running ensembles of simulations
with slightly different initial conditions, designed explicitly to represent
internal variability, or can be estimated on the basis of long control
runs where external forcings are held constant. In the case of both
multi-model and perturbed physics ensembles (see below), there is an
implicit perturbation in the initial state of each run considered, which
means that these ensembles sample both modelling uncertainty and
internal variability jointly.
The ability of models to mimic nature is achieved by simplification
choices that can vary from model to model in terms of the fundamental
numeric and algorithmic structures, forms and values of parameteriza-
tions, and number and kinds of coupled processes included. Simplifi-
cations and the interactions between parameterized and resolved pro-
cesses induce ‘errors’ in models, which can have a leading-order impact
on projections. It is possible to characterize the choices made when
building and running models into structural—indicating the numerical
techniques used for solving the dynamical equations, the analytic form
of parameterization schemes and the choices of inputs for fixed or var-
ying boundary conditions—and parametric—indicating the choices
made in setting the parameters that control the various components
of the model. The community of climate modellers has regularly col-
laborated in producing coordinated experiments forming multi-model
ensembles (MMEs), using both global and regional model families, for
example, CMIP3/5 (Meehl et al., 2007a), ENSEMBLES (Johns et al.,
2011) and Chemistry–Climate Model Validation 1 and 2 (CCM-Val-1
and 2; Eyring et al., 2005), through which structural uncertainty can be
at least in part explored by comparing models, and perturbed physics
ensembles (PPEs, with e.g., Hadley Centre Coupled Model version 3
(HadCM3; Murphy et al., 2004), Model for Interdiciplinary Research On
Climate (MIROC; Yokohata et al., 2012), Community Climate System
Model 3 (CCSM3; Jackson et al., 2008; Sanderson, 2011)), through
which uncertainties in parameterization choices can be assessed in a
given model. As noted below, neither MMEs nor PPEs represent an
adequate sample of all the possible choices one could make in building
a climate model. Also, current models may exclude some processes that
could turn out to be important for projections (e.g., methane clathrate
release) or produce a common error in the representation of a particu-
lar process. For this reason, it is of critical importance to distinguish
two different senses in which the uncertainty terminology is used or
misused in the literature (see also Sections 1.4.2, 9.2.2, 9.2.3, 11.2.1
and 11.2.2). A narrow interpretation of the concept of model uncer-
tainty often identifies it with the range of responses of a model ensem-
ble. In this chapter this type of characterization is referred as model
range or model spread. A broader concept entails the recognition of a
fundamental uncertainty in the representation of the real system that
these models can achieve, given their necessary approximations and
the limits in the scientific understanding of the real system that they
encapsulate. When addressing this aspect and characterizing it, this
chapter uses the term model uncertainty.
The relative role of the different sources of uncertainty—model, sce-
nario and internal variability—as one moves from short- to mid- to
long-term projections and considers different variables at different
spatial scales has to be recognized (see Section 11.3). The three sourc-
es exchange relevance as the time horizon, the spatial scale and the
variable change. In absolute terms, internal variability is generally
estimated, and has been shown in some specific studies (Hu et al.,
2012) to remain approximately constant across the forecast horizon,
with model ranges and scenario/forcing variability increasing over
time. For forecasts of global temperatures after mid-century, scenario
and model ranges dominate the amount of variation due to internally
generated variability, with scenarios accounting for the largest source
1040
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
of uncertainty in projections by the end of the century. For global aver-
age precipitation projections, scenario uncertainty has a much smaller
role even by the end of the 21st century and model range maintains
the largest share across all projection horizons. For temperature and
precipitation projections at smaller spatial scales, internal variability
may remain a significant source of uncertainty up until middle of the
21st century in some regions (Hawkins and Sutton, 2009, 2011; Rowell,
2012; Knutti and Sedláček, 2013). Within single model experiments,
the persistently significant role of internally generated variability for
regional projections even beyond short- and mid-term horizons has
been documented by analyzing relatively large ensembles sampling
initial conditions (Deser et al., 2012a, 2012b).
12.2.3 From Ensembles to Uncertainty Quantification
Ensembles like CMIP5 do not represent a systematically sampled
family of models but rely on self-selection by the modelling groups.
This opportunistic nature of MMEs has been discussed, for example, in
Tebaldi and Knutti (2007) and Knutti et al. (2010a). These ensembles are
therefore not designed to explore uncertainty in a coordinated manner,
and the range of their results cannot be straightforwardly interpreted
as an exhaustive range of plausible outcomes, even if some studies
have shown how they appear to behave as well calibrated probabil-
istic forecasts for some large-scale quantities (Annan and Hargreaves,
2010). Other studies have argued instead that the tail of distributions
is by construction undersampled (Räisänen, 2007). In general, the dif-
ficulty in producing quantitative estimates of uncertainty based on
multiple model output originates in their peculiarities as a statistical
sample, neither random nor systematic, with possible dependencies
among the members (Jun et al., 2008; Masson and Knutti, 2011; Pen-
nell and Reichler, 2011; Knutti et al., 2013) and of spurious nature, that
is, often counting among their members models with different degrees
of complexities (different number of processes explicitly represented or
parameterized) even within the category of general circulation models.
Agreement between multiple models can be a source of information in
an uncertainty assessment or confidence statement. Various methods
have been proposed to indicate regions where models agree on the
projected changes, agree on no change or disagree. Several of those
methods are compared in Box 12.1. Many figures use stippling or
hatching to display such information, but it is important to note that
confidence cannot be inferred from model agreement alone.
Perturbed physics experiments (PPEs) differ in their output interpret-
ability for they can be, and have been, systematically constructed
and as such lend themselves to a more straightforward treatment
through statistical modelling (Rougier, 2007; Sanso and Forest, 2009).
Uncertain parameters in a single model to whose values the output
is known to be sensitive are targeted for perturbations. More often
it is the parameters in the atmospheric component of the model that
are varied (Collins et al., 2006a; Sanderson et al., 2008), and to date
have in fact shown to be the source of the largest uncertainties in
large-scale response, but lately, with much larger computing power
expense, also parameters within the ocean component have been per-
turbed (Collins et al., 2007; Brierley et al., 2010). Parameters in the
land surface schemes have also been subject to perturbation studies
(Fischer et al., 2011; Booth et al., 2012; Lambert et al., 2012). Ranges
of possible values are explored and often statistical models that fit the
relationship between parameter values and model output, that is, emu-
lators, are trained on the ensemble and used to predict the outcome
for unsampled parameter value combinations, in order to explore the
parameter space more thoroughly that would otherwise be computa-
tionally affordable (Rougier et al., 2009). The space of a single model
simulations (even when filtered through observational constraints) can
show a large range of outcomes for a given scenario (Jackson et al.,
2008). However, multi-model ensembles and perturbed physics ensem-
bles produce modes and distributions of climate responses that can
be different from one another, suggesting that one type of ensemble
cannot be used as an analogue for the other (Murphy et al., 2007;
Sanderson et al., 2010; Yokohata et al., 2010; Collins et al., 2011).
Many studies have made use of results from these ensembles to charac-
terize uncertainty in future projections, and these will be assessed and
their results incorporated when describing specific aspects of future
climate responses. PPEs have been uniformly treated across the differ-
ent studies through the statistical framework of analysis of computer
experiments (Sanso et al., 2008; Rougier et al., 2009; Harris et al., 2010)
or, more plainly, as a thorough exploration of alternative responses
reweighted by observational constraints (Murphy et al., 2004; Piani et
al., 2005; Forest et al., 2008; Sexton et al., 2012). In all cases the con-
struction of a probability distribution is facilitated by the systematic
nature of the experiments. MMEs have generated a much more diver-
sified treatment (1) according to the choice of applying weights to the
different models on the basis of past performance or not (Weigel et al.,
2010) and (2) according to the choice between treating the different
models and the truth as indistinguishable or treating each model as
a version of the truth to which an error has been added (Annan and
Hargreaves, 2010; Sanderson and Knutti, 2012). Many studies can be
classified according to these two criteria and their combination, but
even within each of the four resulting categories different studies pro-
duce different estimates of uncertainty, owing to the preponderance
of a priori assumptions, explicitly in those studies that approach the
problem through a Bayesian perspective, or only implicit in the choice
of likelihood models, or weighting. This makes the use of probabilistic
and other results produced through statistical inference necessarily
dependent on agreeing with a particular set of assumptions (Sansom
et al., 2013), given the lack of a full exploration of the robustness of
probabilistic estimates to varying these assumptions.
In summary, there does not exist at present a single agreed on and
robust formal methodology to deliver uncertainty quantification esti-
mates of future changes in all climate variables (see also Section 9.8.3
and Stephenson et al., 2012). As a consequence, in this chapter, state-
ments using the calibrated uncertainty language are a result of the
expert judgement of the authors, combining assessed literature results
with an evaluation of models demonstrated ability (or lack thereof)
in simulating the relevant processes (see Chapter 9) and model con-
sensus (or lack thereof) over future projections. In some cases when a
significant relation is detected between model performance and relia-
bility of its future projections, some models (or a particular parametric
configuration) may be excluded (e.g., Arctic sea ice; Section 12.4.6.1
and Joshi et al., 2010) but in general it remains an open research ques-
tion to find significant connections of this kind that justify some form
of weighting across the ensemble of models and produce aggregated
1041
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Box 12.1 | Methods to Quantify Model Agreement in Maps
The climate change projections in this report are based on ensembles of climate models. The ensemble mean is a useful quantity to
characterize the average response to external forcings, but does not convey any information on the robustness of this response across
models, its uncertainty and/or likelihood or its magnitude relative to unforced climate variability. In the IPCC AR4 WGI contribution
(IPCC, 2007) several criteria were used to indicate robustness of change, most prominently in Figure SPM.7. In that figure, showing
projected precipitation changes, stippling marked regions where at least 90% of the CMIP3 models agreed on the sign of the change.
Regions where less than 66% of the models agreed on the sign were masked white. The resulting large white area was often misin-
terpreted as indicating large uncertainties in the different models’ response to external forcings, but recent studies show that, for the
most part, the disagreement in sign among models is found where projected changes are small and still within the modelled range
of internal variability, that is, where a response to anthropogenic forcings has not yet emerged locally in a statistically significant way
(Tebaldi et al., 2011; Power et al., 2012).
A number of methods to indicate model robustness, involving an assessment of the significance of the change when compared to inter-
nal variability, have been proposed since AR4. The different methods share the purpose of identifying regions with large, significant or
robust changes, regions with small changes, regions where models disagree or a combination of those. They do, however, use different
assumptions about the statistical properties of the model ensemble, and therefore different criteria for synthesizing the information
from it. Different methods also differ in the way they estimate internal variability. We briefly describe and compare several of these
methods here.
Method (a): The default method used in Chapters 11,12 and 14 as well as in the Annex I (hatching only) is shown in Box 12.1, Figure
1a, and is based on relating the climate change signal to internal variability in 20-year means of the models as a reference
3
. Regions
where the multi-model mean change exceeds two standard deviations of internal variability and where at least 90% of the models
agree on the sign of change are stippled and interpreted as ‘large change with high model agreement’. Regions where the model mean
is less than one standard deviation of internal variability are hatched and interpreted as ‘small signal or low agreement of models’. This
can have various reasons: (1) changes in individual models are smaller than internal variability, or (2) although changes in individual
models are significant, they disagree about the sign and the multi-model mean change remains small. Using this method, the case
where all models scatter widely around zero and the case where all models agree on near zero change therefore are both hatched
(e.g., precipitation change over the Amazon region by the end of the 21st century, which the following methods mark as ‘inconsistent
model response’).
Method (b): Method (a) does not distinguish the case where all models agree on no change and the case where, for example, half of
the models show a significant increase and half a decrease. The distinction may be relevant for many applications and a modification
of method (a) is to restrict hatching to regions where there is high agreement among the models that the change will be ‘small’, thus
eliminating the ambiguous interpretation ‘small or low agreement’ in (a). In contrast to method (a) where the model mean is com-
pared to variability, this case (b) marks regions where at least 80% of the individual models show a change smaller than two standard
deviations of variability with hatching. Grid points where many models show significant change but don’t agree are no longer hatched
(Box 12.1, Figure 1b).
Method (c): Knutti and Sedláček (2013) define a dimensionless robustness measure, R, which is inspired by the signal-to-noise ratio
and the ranked probability skill score. It considers the natural variability and agreement on magnitude and sign of change. A value of
R = 1 implies perfect model agreement; low or negative values imply poor model agreement (note that by definition R can assume any
negative value). Any level of R can be chosen for the stippling. For illustration, in Box 12.1, Figure 1c, regions with R > 0.8 are marked
with small dots, regions with R > 0.9 with larger dots and are interpreted as ‘robust large change’. This yields similar results to method
(a) for the end of the century, but with some areas of moderate model robustness (R > 0.8) already for the near-term projections,
even though the signal is still within the noise. Regions where at least 80% of the models individually show no significant change
are hatched and interpreted as ‘changes unlikely to emerge from variability’
4
.There is less hatching in this method than in method (a),
3
The internal variability in this method is estimated using pre-industrial control runs for each of the models which are at least 500 years long. The first 100 years of
the pre-industrial are ignored. Variability is calculated for every grid point as the standard deviation of non-overlapping 20-year means, multiplied by the square
root of 2 to account for the fact that the variability of a difference in means is of interest. A quadratic fit as a function of time is subtracted from these at every grid
point to eliminate model drift. This is by definition the standard deviation of the difference between two independent 20-year averages having the same variance
and estimates the variation of that difference that would be expected due to unforced internal variability. The median across all models of that quantity is used.
4
Variability in methods b–d is estimated from interannual variations in the base period within each model.
(continued on next page)
1042
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
DJF mean precipitation change (RCP8.5)
Box 12.1, Figure 1 | Projected change in December to February precipitation for 2016–2035 and 2081–2100, relative to 1986–2005 from CMIP5 models. The
choice of the variable and time frames is just for illustration of how the different methods compare in cases with low and high signal-to-noise ratio (left and right
column, respectively). The colour maps are identical along each column and only stippling and hatching differ on the basis of the different methods. Different methods
for stippling and hatching are shown determined (a) from relating the model mean to internal variability, (b) as in (a) but hatching here indicates high agreement for
‘small change’, (c) by the robustness measure by Knutti and Sedláček (2013), (d) by the method proposed by Tebaldi et al. (2011) and (e) by the method by Power et
al. (2012). Detailed technical explanations for each method are given in the text. 39 models are used in all panels.
Box 12.1 (continued)
1043
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Box 12.1 (continued)
because it requires 80% of the models to be within variability, not just the model average. Regions where at least 50% of the models
show significant change but R< 0.5 are masked as white to indicate ‘models disagreeing on the projected change projections’ (Box
12.1, Figure 1c).
Method (d): Tebaldi et al. (2011) start from IPCC AR4 SPM7 but separate lack of model agreement from lack of signal (Box 12.1, Figure
1e). Grid points are stippled and interpreted as ‘robust large change’ when more than 50% of the models show significant change and
at least 80% of those agree on the sign of change. Grid points where more than 50% of the models show significant change but less
than 80% of those agree on the sign of change are masked as white and interpreted as ‘unreliable’. The results are again similar to
the methods above. No hatching was defined in that method (Box 12.1 Figure 1d). (See also Neelin et al., 2006 for a similar approach
applied to a specific regional domain.)
Method (e): Power et al. (2012) identify three distinct regions using various methods in which projections can be very loosely described
as either: ‘statistically significant’, ‘small (relative to temporal variability) or zero, but not statistically significant’ or ‘uncertain’. The
emphasis with this approach is to identify robust signals taking the models at face value and to address the questions: (1) What
will change? (2) By how much? and (3) What will not change? The underlying consideration here is that statistical testing under the
assumption of model independence provides a worthwhile, albeit imperfect, line of evidence that needs to be considered in conjunction
with other evidence (e.g., degree of interdependence, ability of models to simulate the past), in order to assess the degree of confidence
one has in a projected change.
The examples given here are not exhaustive but illustrate the main ideas. Other methods include simply counting the number of models
agreeing on the sign (Christensen et al., 2007), or varying colour hue and saturation to indicate magnitude of change and robustness
of change separately (Kaye et al., 2012). In summary, there are a variety of ways to characterize magnitude or significance of change,
and agreement between models. There is also a compromise to make between clarity and richness of information. Different methods
serve different purposes and a variety of criteria can be justified to highlight specific properties of multi-model ensembles. Clearly only
a subset of information regarding robust and uncertain change can be conveyed in a single plot. The methods above convey some
important pieces of this information, but obviously more information could be provided if more maps with additional statistics were
provided. In fact Annex I provides more explicit information on the range of projected changes evident in the models (e.g., the median,
and the upper and lower quartiles). For most of the methods there is a necessity to choose thresholds for the level of agreement that
cannot be identified objectively, but could be the result of individual, application-specific evaluations. Note also that all of the above
methods measure model agreement in an ensemble of opportunity, and it is impossible to derive a confidence or likelihood statement
from the model agreement or model spread alone, without considering consistency with observations, model dependence and the
degree to which the relevant processes are understood and reflected in the models (see Section 12.2.3).
The method used by Power et al. (2012) differs from the other methods in that it tests the statistical significance of the ensemble mean
rather than a single simulation. As a result, the area where changes are significant increases with an increasing number of models.
Already for the period centred on 2025, most of the grid points when using this method show significant change in the ensemble
mean whereas in the other methods projections for this time period are classified as changes not exceeding internal variability. The
reason is that the former produces a statement about the mean of the distribution being significantly different from zero, equivalent to
treating the ensemble as ‘truth plus error’, that is, assuming that the models are independent and randomly distributed around reality.
Methods a–d, on the other hand, use an ‘indistinguishable’ interpretation, in which each model and reality are drawn from the same
distribution. In that case, the stippling and hatching characterize the likelihood of a single member being significant or not, rather than
the ensemble mean. There is some debate in the literature on how the multi-model ensembles should be interpreted statistically. This
and past IPCC reports treat the model spread as some measure of uncertainty, irrespective of the number of models, which implies an
‘indistinguishable’ interpretation. For a detailed discussion readers are referred to the literature (Tebaldi and Knutti, 2007; Annan and
Hargreaves, 2010; Knutti et al., 2010a, 2010b; Annan and Hargreaves, 2011a; Sanderson and Knutti, 2012).
1044
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
future projections that are significantly different from straightforward
one model–one vote (Knutti, 2010) ensemble results. Therefore, most
of the analyses performed for this chapter make use of all available
models in the ensembles, with equal weight given to each of them
unless otherwise stated.
12.2.4 Joint Projections of Multiple Variables
While many of the key processes relevant to the simulation of single
variables are understood, studies are only starting to focus on assess-
ing projections of joint variables, especially when extremes or varia-
bility in the individual quantities are of concern. A few studies have
addressed projected changes in joint variables, for example, by combin-
ing mean temperature and precipitation (Williams et al., 2007; Tebaldi
and Lobell, 2008; Tebaldi and Sanso, 2009; Watterson, 2011; Watter-
son and Whetton, 2011a; Sexton et al., 2012), linking soil moisture,
precipitation and temperature mean and variability (Seneviratne et al.,
2006; Fischer and Schär, 2009; Koster et al., 2009b, 2009c), combining
temperature and humidity (Diffenbaugh et al., 2007; Fischer and Schär,
2010; Willett and Sherwood, 2012), linking summertime temperature
and soil moisture to prior winter snowpack (Hall et al., 2008) or linking
precipitation change to circulation, moisture and moist static energy
budget changes (Neelin et al., 2003; Chou and Neelin, 2004; Chou et
al., 2006, 2009). Models may have difficulties simulating all relevant
interactions between atmosphere and land surface and the water cycle
that determine the joint response, observations to evaluate models are
often limited (Seneviratne et al., 2010), and model uncertainties are
therefore large (Koster et al., 2006; Boé and Terray, 2008; Notaro, 2008;
Fischer et al., 2011). In some cases, correlations between, for example,
temperature and precipitation or accumulated precipitation and tem-
perature have found to be too strong in climate models (Trenberth and
Shea, 2005; Hirschi et al., 2011). The situation is further complicated
by the fact that model biases in one variable affect other variables.
The standard method for model projections is to subtract model biases
derived from control integrations (assuming that the bias remains con-
stant in a future scenario integration). Several studies note that this
may be problematic when a consistent treatment of biases in multiple
variables is required (Christensen et al., 2008; Buser et al., 2009), but
there is no consensus at this stage for a methodology addressing this
problem (Ho et al., 2012). More generally the existence of structural
errors in models according to which an unavoidable discrepancy (Rou-
gier, 2007) between their simulations and reality cannot be avoided
is relevant here, as well as for univariate projections. In the recent lit-
erature an estimate of this discrepancy has been proposed through
the use of MMEs, using each model in turn as a surrogate for reali-
ty, and measuring the distance between it and the other models of
the ensemble. Some summary statistic of these measures is then used
to estimate the distance between models and the real world (Sexton
and Murphy, 2012; Sexton et al., 2012; Sanderson, 2013). Statistical
frameworks to deal with multivariate projections are challenging even
for just two variables, as they have to address a trade-off between
modelling the joint behavior at scales that are relevant for impacts—
that is, fine spatial and temporal scales, often requiring complex spa-
tio-temporal models—and maintaining computational feasibility. In
one instance (Tebaldi and Sanso, 2009) scales were investigated at
the seasonal and sub-continental level, and projections of the forced
response of temperature and precipitation at those scales did not show
significant correlations, likely because of the heterogeneity of the rela-
tion between the variables within those large averaged regions and
seasons. In Sexton et al. (2012) the spatial scale focussed on regions of
Great Britain and correlation emerged as more significant, for exam-
ple, between summer temperatures and precipitation amounts. Fischer
and Knutti (2013) estimated strong relationships between variables
making up impact relevant indices (e.g., temperature and humidi-
ty) and showed how in some cases, uncertainties across models are
larger for a combined variable than if the uncertainties in the individ-
ual underlying variables were treated independently (e.g., wildfires),
whereas in other cases theuncertainties in the combined variables are
smaller than in the individual ones (e.g., heat stress for humans).
Even while recognizing the need for joint multivariate projections, the
above limitations at this stage prevent a quantitative assessment for
most cases. A few robust qualitative relationships nonetheless emerge
from the literature and these are assessed, where appropriate, in the
rest of the chapter. For applications that are sensitive to relationships
between variables, but still choose to use the multi-model framework
to determine possible ranges for projections, sampling from univari-
ate ranges may lead to unrealistic results when significant correlations
exist. IPCC assessments often show model averages as best estimates,
but such averages can underestimate spatial variability, and more in
general they neither represent any of the actual model states (Knutti et
al., 2010a) nor do they necessarily represent the joint best estimate in a
multivariate sense. Impact studies usually need temporally and spatial-
ly coherent multivariate input from climate model simulations. In those
cases, using each climate model output individually and feeding it into
the impact model, rather than trying to summarise a multivariate distri-
bution from the MME and sample from it, is likely to be more consist-
ent, assuming that the climate model itself correctly captures the spa-
tial covariance, the temporal co-evolution and the relevant feedbacks.
12.3 Projected Changes in Forcing Agents,
Including Emissions and Concentrations
The experiments that form the basis of global future projections dis-
cussed in this chapter are extensions of the simulations of the observa-
tional record discussed in Chapters 9 and 10. The scenarios assessed in
AR5, introduced in Chapter 1, include four new scenarios designed to
explore a wide range of future climate characterized by representative
trajectories of well-mixed greenhouse gas (WMGHG) concentrations
and other anthropogenic forcing agents. These are described further
in Section 12.3.1. The implementation of forcing agents in model pro-
jections, including natural and anthropogenic aerosols, ozone and land
use change are discussed in Section 12.3.2, with a strong focus on
CMIP5 experiments. Global mean emissions, concentrations and RFs
applicable to the historical record simulations assessed in Chapters 8,
9 and 10, and the future scenario simulations assessed here, are listed
in Annex II. Global mean RF for the 21st century consistent with these
scenarios, derived from CMIP5 and other climate model studies, is dis-
cussed in Section 12.3.3.
1045
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
12.3.1 Description of Scenarios
Long-term climate change projections reflect how human activities or
natural effects could alter the climate over decades and centuries. In
this context, defined scenarios are important, as using specific time
series of emissions, land use, atmospheric concentrations or RF across
multiple models allows for coherent climate model intercomparisons
and synthesis. Some scenarios present a simple stylized future (not
accompanied by a socioeconomic storyline) and are used for pro-
cess understanding. More comprehensive scenarios are produced by
Integrated Assessment Models (IAMs) as internally consistent sets of
emissions and socioeconomic assumptions (e.g., regarding population
and socioeconomic development) with the aim of presenting sever-
al plausible future worlds (see Section 1.5.2 and Box 1.1). In general
it is these scenarios that are used for policy relevant climate change,
impact, adaptation and mitigation analysis. It is beyond the scope of
this report to consider the full range of currently published scenarios
and their implications for mitigation policy and climate targets—that
is covered by the Working Group III contribution to the AR5. Here, we
focus on the RCP scenarios used within the CMIP5 intercomparison
exercise (Taylor et al. 2012) along with the SRES scenarios (IPCC, 2000)
developed for the IPCC Third Assessment Report (TAR) but still widely
used by the climate community.
12.3.1.1 Stylized Concentration Scenarios
A 1% per annum compound increase of atmospheric CO
2
concen-
tration until a doubling or a quadrupling of its initial value has been
widely used since the second phase of CMIP (Meehl et al., 2000) and
the Second Assessment Report (Kattenberg et al., 1996). This stylized
scenario is a useful benchmark for comparing coupled model climate
sensitivity, climate feedback and transient climate response, but is not
used directly for future projections. The exponential increase of CO
2
concentration induces approximately a linear increase in RF due to
a ‘saturation effect’ of the strong absorbing bands (Augustsson and
Ramanathan, 1977; Hansen et al., 1988; Myhre et al., 1998). Thus, a
linear ramp function in forcing results from these stylized pathways,
adding to their suitability for comparative diagnostics of the models’
climate feedbacks and inertia. The CMIP5 intercomparison project
again includes such a stylized pathway, in which the CO
2
concentration
reaches twice the initial concentration after 70 years and four times
the initial concentration after 140 years. The corresponding RFs are
3.7 W m
–2
(Ramaswamy et al., 2001) and 7.4 W m
–2
respectively with
a range of ±20% accounting for uncertainties in radiative transfer cal-
culations and rapid adjustments (see Section 8.3.2.1), placing them
within the range of the RFs at the end of the 21st century for the
future scenarios presented below. The CMIP5 project also includes a
second stylized experiment in which the CO
2
concentration is quadru-
pled instantaneously, which allows a distinction between effective RFs
and longer-term climate feedbacks (Gregory et al., 2004).
12.3.1.2 The Socioeconomic Driven Scenarios from the Special
Report on Emission Scenarios
The climate change projections undertaken as part of CMIP3 and dis-
cussed in AR4 were based primarily on the SRES A2, A1B and B1 sce-
narios (IPCC, 2000). These scenarios were developed using IAMs and
resulted from specific socioeconomic scenarios, that is, from storylines
about future demographic and economic development, regionaliza-
tion, energy production and use, technology, agriculture, forestry, and
land use. All SRES scenarios assumed that no climate mitigation policy
would be undertaken. Based on these SRES scenarios, global climate
models were then forced with corresponding WMGHG and aerosol
concentrations, although the degree to which models implemented
these forcings differed (Meehl et al., 2007b, Table 10.1). The result-
ing climate projections, together with the socioeconomic scenarios on
which they are based, have been widely used in further analysis by the
impact, adaptation and vulnerability research communities.
12.3.1.3 The New Concentration Driven Representative
Concentration Pathway Scenarios, and Their Extensions
As introduced in Box 1.1 and mentioned in Section 12.1, a new parallel
process for scenario development was proposed in order to facilitate
the interactions between the scientific communities working on cli-
mate change, adaptation and mitigation (Hibbard et al., 2007; Moss et
al., 2008, 2010; van Vuuren et al., 2011). These new scenarios, Repre-
sentative Concentration Pathways, are referred to as pathways in order
to emphasize that they are not definitive scenarios, but rather inter-
nally consistent sets of time-dependent forcing projections that could
potentially be realized with more than one underlying socioeconomic
scenario. The primary products of the RCPs are concentrations but they
also provide gas emissions. They are representative in that they are one
of several different scenarios, sampling the full range of published sce-
narios (including mitigation scenarios) at the time they were defined,
that have similar RF and emissions characteristics. They are identified
by the approximate value of the RF (in W m
–2
) at 2100 or at stabiliza-
tion after 2100 in their extensions, relative to pre-industrial (Moss et
al., 2008; Meinshausen et al., 2011c). RCP2.6 (the lowest of the four,
also referred to as RCP3-PD) peaks at 3.0 W m
–2
and then declines to
2.6 W m
–2
in 2100, RCP4.5 (medium-low) and RCP6.0 (medium-high)
stabilize after 2100 at 4.2 and 6.0 W m
–2
respectively, while RCP8.5
(highest) reaches 8.3 W m
–2
in 2100 on a rising trajectory (see also
Figure 12.3a which takes into account the efficacies of the various
anthropogenic forcings). The primary objective of these scenarios is to
provide all the input variables necessary to run comprehensive climate
models in order to reach a target RF (Figure 12.2). These scenarios
were developed using IAMs that provide the time evolution of a large
ensemble of anthropogenic forcings (concentration and emission of
gas and aerosols, land use changes, etc.) and their individual RF values
(Moss et al., 2008, 2010; van Vuuren et al., 2011). Note that due to the
substantial uncertainties in RF, these forcing values should be under-
stood as comparative ‘labels’, not as exact definitions of the forcing
that is effective in climate models. This is because concentrations or
emissions, rather than the RF itself, are prescribed in the CMIP5 climate
model runs. The forcing as manifested in climate models is discussed
in Section 12.3.3.
Various steps were necessary to turn the selected ‘raw’ RCP scenarios
from the IAMs into data sets usable by the climate modelling commu-
nity. First, harmonization with historical data was performed for emis-
sions of reactive gases and aerosols (Lamarque et al., 2010; Granier
et al., 2011; Smith et al., 2011), land use (Hurtt et al., 2011), and for
GHG emissions and concentrations (Meinshausen et al., 2011c). Then
1046
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
2000 2050 2100 2150220022502300
Year
0
2
4
6
8
10
12
Anthropogenic Radiative Forcing (W m
-2
)
SRES-B1
SRES-A1B
SRES-A2
RCP2.6
RCP4.5
RCP6.0
RCP8.5
LU
aerosolother
O
3
N
2
OCH
4
CO
2
RCP8.5 (2100)
RCP6.0 (2100)
RCP4.5 (2100)
RCP2.6 (2100)
present day (2010)
(W m
2
)
Contribution of individual forcings to the total
−1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
LU
aerosolother
O
3
N
2
OCH
4
CO
2
RCP8.5 (2100)
RCP6.0 (2100)
RCP4.5 (2100)
RCP2.6 (2100)
present day (2010)
(%)
Contribution of individual forcings as a percentage of the total
−60 −40 −200 20 40 60 80 100
b)
c)
a)
GHG
GHG
atmospheric chemistry runs were performed to estimate ozone and
aerosol distributions (Lamarque et al., 2011). Finally, a single carbon
cycle model with a representation of carbon–climate feedbacks was
used in order to provide consistent values of CO
2
concentration for
the CO
2
emission provided by a different IAM for each of the scenari-
os. This methodology was used to produce consistent data sets across
scenarios but does not provide uncertainty estimates for them. After
these processing steps, the final RCP data sets comprise land use
data, harmonized GHG emissions and concentrations, gridded reactive
gas and aerosol emissions, as well as ozone and aerosol abundance
fields. These data are used as forcings in individual climate models. The
number and type of forcings included primarily depend on the exper-
iment. For instance, while the CO
2
concentration is prescribed in most
experiments, CO
2
emissions are prescribed in some others (see Box 6.4
and Section 12.3.2.1). Which of these forcings are included in individ-
ual CMIP5 models, and variations in their implementation, is described
in Section 12.3.2.2.
During this development process, the total RF and the RF of individual
forcing agents have been estimated by the IAMs and made availa-
ble via the RCP database (Meinshausen et al., 2011c). Each individual
anthropogenic forcing varies from one scenario to another. They have
Figure 12.3 | (a) Time evolution of the total anthropogenic (positive) and anthropogenic aerosol (negative) radiative forcing (RF) relative to pre-industrial (about 1765) between
2000 and 2300 for RCP scenarios and their extensions (continuous lines), and SRES scenarios (dashed lines) as computed by the Integrated Assessment Models (IAMs) used to
develop those scenarios. The four RCP scenarios used in CMIP5 are: RCP2.6 (dark blue), RCP4.5 (light blue), RCP6.0 (orange) and RCP8.5 (red). The three SRES scenarios used
in CMIP3 are: B1 (blue, dashed), A1B (green, dashed) and A2 (red, dashed). Positive values correspond to the total anthropogenic RF. Negative values correspond to the forcing
from all anthropogenic aerosol–radiation interactions (i.e., direct effects only). The total RF of the SRES and RCP families of scenarios differs in 2000 because the number of forc-
ings represented and our knowledge about them have changed since the TAR. The total RF of the RCP family is computed taking into account the efficacy of the various forcings
(Meinshausen et al., 2011a). (b) Contribution of the individual anthropogenic forcings to the total RF in year 2100 for the four RCP scenarios and at present day (year 2010). The
individual forcings are gathered into seven groups: carbon dioxide (CO
2
), methane (CH
4
), nitrous oxide (N
2
O), ozone (O
3
), other greenhouse gases, aerosol (all effects unlike in (a),
i.e., aerosol–radiation and aerosol–cloud interactions, aerosol deposition on snow) and land use (LU). (c) As in (b), but the individual forcings are relative to the total RF (i.e., RF
x
/
RF
tot
, in %, with RF
x
individual RFs and RF
tot
total RF). Note that the RFs in (b) and (c) are not efficacy adjusted, unlike in (a). The values shown in (a) are summarized in Table AII.6.8.
The values shown in (b) and (c) have been directly extracted from data files (hosted at http://tntcat.iiasa.ac.at:8787/RcpDb/) compiled by the four modelling teams that developed
the RCP scenarios and are summarized in Tables AII.6.1 to AII.6.3 for CO
2
, CH
4
and N
2
O respectively.
1047
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
been aggregated into a few groups in Figure 12.3b and c. The total
anthropogenic RF estimated by the IAMs in 2010 is about 0.15 W m
–2
lower than Chapter 8’s best estimate of ERF in 2010 (2.2 W m
–2
), the
difference arising from a revision of the RF due to aerosols and land
use in the current assessment compared to AR4. All the other individ-
ual forcings are consistent to within 0.02 W m
–2
. The change in CO
2
concentration is the main cause of difference in the total RF among
the scenarios (Figure 12.3b). The relative contribution
5
of CO
2
to the
total anthropogenic forcing is currently (year 2010) about 80 to 90%
and does not vary much across the scenarios (Figure 12.3c), as was
also the case for SRES scenarios (Ramaswamy et al., 2001). Aerosols
have a large negative contribution to the total forcing (about –40 to
–50% in 2010), but this contribution decreases (in both absolute and
relative terms) in the future for all the RCPs scenarios. This means that
while anthropogenic aerosols have had a cooling effect in the past,
their decrease in all RCP scenarios relative to current levels is expected
to have a net warming effect in the future (Levy II et al., 2013; see also
Figure 8.20). The 21st century decrease in the magnitude of future aer-
osol forcing was not as large and as rapid in the SRES scenarios (Figure
12.3a). However, even in the SRES scenarios, aerosol effects were
expected to have a diminishing role in the future compared to GHG
forcings, mainly because of the accumulation of GHG in the atmos-
phere (Dufresne et al., 2005). Other forcings do not change much in
the future, except CH
4
which increases in the RCP8.5 scenario. Note
that the estimates of all of these individual RFs are subject to many
uncertainties (see Sections 7.5, 8.5 and 11.3.6). In this section and in
Table AII.6.8, the RF values for RCP scenarios are derived from pub-
lished equivalent-CO
2
(CO
2
eq) concentration data that aggregates all
anthropogenic forcings including GHGs and aerosols. The conversion
to RF uses the formula: RF = 3.71/ln(2) ∙ ln(CO
2
eq/278) W m
–2
, where
CO
2
eq is in ppmv.
The four RCPs (Meinshausen et al., 2011c) are based on IAMs up to the
end of the 21st century only. In order to investigate longer-term climate
change implications, these RCPs were also extended until 2300. The
extensions, formally named Extended Concentration Pathways (ECPs)
but often simply referred to as RCP extensions, use simple assump-
tions about GHG and aerosol emissions and concentrations beyond
2100 (such as stabilization or steady decline) and were designed as
hypothetical ‘what-if’ scenarios, not as an outcome of an IAM assum-
ing socioeconomic considerations beyond 2100 (Meinshausen et al.,
2011c) (see Box 1.1). In order to continue to investigate a broad range
of possible climate futures, RCP2.6 assumes small constant net nega-
tive emissions after 2100 and RCP8.5 assumes stabilization with high
emissions between 2100 and 2150, then a linear decrease until 2250.
The two middle RCPs aim for a smooth stabilization of concentrations
by 2150. RCP8.5 stabilizes concentrations only by 2250, with CO
2
concentrations of approximately 2000 ppmv, nearly seven times the
pre-industrial level. As RCP2.6 implies net negative CO
2
emissions after
around 2070 and throughout the extension, CO
2
concentrations slowly
reduce towards 360 ppmv by 2300.
5
The range of the relative contribution of CO
2
and aerosols to the total anthropogenic forcing is derived here from the RF values given by the IAMs and the best estimate assessed
in Chapter 8.
12.3.1.4 Comparison of Special Report on Emission Scenarios
and Representative Concentration Pathway Scenarios
The four RCP scenarios used in CMIP5 lead to RF values that range from
2.3 to 8.0 W m
–2
at 2100, a wider range than that of the three SRES
scenarios used in CMIP3 which vary from 4.2 to 8.1 W m
–2
at 2100 (see
Table AII.6.8 and Figure 12.3). The SRES scenarios do not assume any
policy to control climate change, unlike the RCP scenarios. The RF of
RCP2.6 is hence lower by 1.9 W m
–2
than the three SRES scenarios and
very close to the ENSEMBLES E1 scenario (Johns et al., 2011). RCP4.5
and SRES B1 have similar RF at 2100, and comparable time evolution
(within 0.2 W m
–2
). The RF of SRES A2 is lower than RCP8.5 through-
out the 21st century, mainly due to a faster decline in the radiative
effect of aerosols in RCP8.5 than SRES A2, but they converge to within
0.1 W m
–2
at 2100. RCP6.0 lies in between SRES B1 and SRES A1B.
Results obtained with one General Circulation Model (GCM) (Dufresne
et al., 2013) and with a reduced-complexity model (Rogelj et al., 2012)
confirm that the differences in temperature responses are consistent
with the differences in RFs estimates. RCP2.6, which assumes strong
mitigation action, yields a smaller temperature increase than any SRES
scenario. The temperature increase with the RCP4.5 and SRES B1 sce-
narios are close and the temperature increase is larger with RCP8.5
than with SRES A2. The spread of projected global mean temperature
for the RCP scenarios (Section 12.4.1) is considerably larger (at both
the high and low response ends) than for the three SRES scenarios
used in CMIP3 (B1, A1B and A2) as a direct consequence of the larger
range of RF across the RCP scenarios compared to that across the
three SRES scenarios (see analysis of SRES versus RCP global tempera-
ture projections in Section 12.4.9 and Figure 12.40).
12.3.2 Implementation of Forcings in Coupled Model
Intercomparison Project Phase 5 Experiments
The CMIP5 experimental protocol for long-term transient climate
experiments prescribes a common basis for a comprehensive set of
anthropogenic forcing agents acting as boundary conditions in three
experimental phases—historical, RCPs and ECPs (Taylor et al., 2012).
To permit common implementations of this set of forcing agents in
CMIP5 models, self-consistent forcing data time series have been com-
puted and provided to participating models (see Sections 9.3.2.2 and
12.3.1.3) comprising emissions or concentrations of GHGs and related
compounds, ozone and atmospheric aerosols and their chemical pre-
cursors, and land use change.
The forcing agents implemented in Atmosphere–Ocean General Cir-
culation Models (AOGCMs) and ESMs used to make long-term cli-
mate projections in CMIP5 are summarized in Table 12.1. The number
of CMIP5 models listed here is about double the number of CMIP3
models listed in Table 10.1 of AR4 (Meehl et al., 2007b).
Natural forcings (arising from solar variability and aerosol emissions
via volcanic activity) are also specified elements in the CMIP5 exper-
imental protocol, but their future time evolutions are not prescribed
1048
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Table 12.1 | Radiative forcing agents in the CMIP5 multi-model global climate projections. See Table 9.A.1 for descriptions of the models and main model references. Earth System Models (ESMs) are highlighted in bold. Numeric superscripts
indicate model-specific references that document forcing implementations. Forcing agents are mostly implemented in close conformance with standard prescriptions (Taylor et al., 2012) and recommended data sets (Lamarque et al., 2010;
Cionni et al., 2011; Lamarque et al., 2011; Meinshausen et al., 2011c) provided for CMIP5. Variations in forcing implementations are highlighted with superscripts and expanded in the table footnotes. Entries mean: n.a.: Forcing agent is not
included in either the historical or future scenario simulations; Y: Forcing agent included (via prescribed concentrations, distributions or time series data); E: Concentrations of forcing agent calculated interactively driven by prescribed emissions
or precursor emissions; Es: Concentrations of forcing agent calculated interactively constrained by prescribed surface concentrations. For a more detailed classification of ozone chemistry and ozone forcing implementations in CMIP5 models
see Eyring et al. (2013).
Model
Forcing Agents
Greenhouse Gases Aerosols Other
CO
2
ce
CH
4
N
2
O Trop O
3
Strat O
3
CFCs SO
4
Black
carbon
Organic
carbon
Nitrate
Cloud
albedo
effect
ac
Cloud
lifetime
effect
ac
Dust Volcanic Sea salt
Land
use
Solar
ACCESS-1.0
1
Y
p
Y Y Y
b
Y
b
Y E E E n.a. Y Y Y
pd
Y
v5
Y
pd
n.a. Y
ACCESS-1.3
1
Y
p
Y Y Y
b
Y
b
Y E E E n.a. Y Y n.a. Y
v5
Y
pd
n.a. Y
BCC-CSM1.1
2
Y/E
p
Y Y Y
b
Y
b
Y Y
a
Y
a
Y
a
n.a. n.a. n.a. Y
a
Y
v0
Y
a
n.a. Y
BCC-CSM1.1(m)
2
Y/E
p
Y Y Y
b
Y
b
Y Y
a
Y
a
Y
a
n.a. n.a. n.a. Y
a
Y
v0
Y
a
n.a. Y
BNU-ESM Y/E
p
Y Y Y
a
Y
a
Y Y
a
Y
a
Y
a
n.a. n.a. n.a. Y
a
Y
v0
Y
a
n.a. Y
CanCM4 Y Y Y Y
b
Y
b
Y E E E n.a. Y
so
n.a. Y
pd
Y/E
st,v0
Y
pd
n.a. Y
CanESM2 Y/E
p
Y Y Y
b
Y
b
Y E E E n.a. Y
so
n.a. Y
pd
Y/E
st,v0
Y
pd
Y
cr
Y
CCSM4
3
Y
p
Y Y Y
a
Y
a
Y Y
a
Y
a
Y
a
n.a. n.a. n.a. Y
a
Y
v0
Y
a
Y Y
CESM1(BGC)
4
Y/E
p
Y Y Y
a
Y
a
Y Y
a
Y
a
Y
a
n.a. n.a. n.a. Y
a
Y
v0
Y
a
Y Y
CESM1(CAM5)
5
Y
p
Y Y Y
a
Y
a
Y E E E n.a. Y Y E Y
v0
E Y Y
CESM1(CAM5.1,FV2)
5
Y
p
Y Y Y
a
Y
a
Y E E E n.a. Y Y E Y
v0
E Y Y
CESM1(FASTCHEM) Y
p
Y
a
Y E E Y E Y
a
Y
a
n.a. n.a. n.a. Y
a
Y
v0
Y
a
Y Y
CESM1(WACCM)
6
Es
p
Es Es E/Es
op
E/Es
op
Es Y Y Y n.a. n.a. n.a. Y
a
Y
v0
Y
a
Y Y
CMCC-CESM
7
Y Y Y Y
b
Y
b
Y Y
a
n.a. n.a. n.a. Y
so
n.a. Y
fx
n.a. Y
fx
n.a. Y
or
CMCC-CM Y Y Y Y
b
Y
b
Y Y
a
n.a. n.a. n.a. Y
so
n.a. Y
fx
n.a. Y
fx
n.a. Y
or
CMCC-CMS Y Y Y Y
b
Y
b
Y Y
a
n.a. n.a. n.a. Y
so
n.a. Y
fx
n.a. Y
fx
n.a. Y
or
CNRM-CM5
8
Y Y Y Y
c
Y
c
Y Y
e
Y
e
Y
e
n.a. Y
so,ic
n.a. Y
e
Y
v1
Y
e
n.a. Y
CSIRO-Mk3.6.0
9
Y Y Y Y
b
Y
b
Y E E E n.a. Y Y Y
pd
Y
v0
Y
pd
n.a. Y
EC-EARTH
10
Y Y Y Y
b
Y
b
Y Y
a
Y
a
Y
a
n.a. n.a. n.a. Y
a
Y
v1
Y
a
Y Y
FGOALS-g2
11
Y Y Y Y
b
Y
b
Y Y
a
Y
a
Y
a
n.a. Y Y Y
a
n.a. Y
a
n.a. Y
FGOALS-s2
12
Y/E Y Y Y
b
Y
b
Y Y
a
Y
a
Y
a
n.a. n.a. n.a. Y
a
Y
v0
Y
a
n.a. Y
FIO-ESM Y/E Y Y Y
a
Y
a
Y Y
a
Y
a
Y
a
n.a. n.a. n.a. Y
a
Y
v0
Y
a
n.a. Y
GFDL-CM3
13
Y
p
Y/Es
rc
Y/Es
rc
E E Y/Es
rc
E E E n.a./E
rc
Y Y E
pd
Y/E
st,v0
E
pd
Y Y
GFDL-ESM2G Y/E
p
Y Y Y
b
Y
b
Y Y
a
Y
a
Y
a
n.a. n.a. n.a. Y
fx
Y
v0
Y
fx
Y Y
GFDL-ESM2M Y/E
p
Y Y Y
b
Y
b
Y Y
a
Y
a
Y
a
n.a. n.a. n.a. Y
fx
Y
v0
Y
fx
Y Y
GISS-E2-p1
14
Y Y Y Y
d
Y
d
Y Y Y Y Y Y n.a. Y
fx
Y
v4
Y
fx
Y Y
or
GISS-E2-p2
14
Y Es/E
hf
Es E E Es/E
hf
E E E E Y n.a. Y
pd
Y
v4
Y
pd
Y Y
or
(continued on next page)
1049
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Model
Forcing Agents
Greenhouse Gases Aerosols Other
CO
2
ce
CH
4
N
2
O Trop O
3
Strat O
3
CFCs SO
4
Black
carbon
Organic
carbon
Nitrate
Cloud
albedo
effect
ac
Cloud
lifetime
effect
ac
Dust Volcanic Sea salt
Land
use
Solar
GISS-E2-p3
14
Y Es/E
hf
Es E E Es/E
hf
E E E E Y n.a. Y
pd
Y
v4
Y
pd
Y Y
or
HadCM3 Y
p
Y Y Y
b
Y
b
Y E n.a. n.a. n.a. Y
so
n.a. n.a. Y
v2
n.a. n.a. Y
HadGEM2-AO
15
Y
p
Y Y Y
b
Y
b
Y E E E n.a. Y Y Y
pd
Y
v2
Y
pd
Y Y
HadGEM2-CC
16,17
Y
p
Y Y Y
b
Y
b
Y E E E n.a. Y Y Y
pd
Y
v2
Y
pd
Y Y
HadGEM2-ES
16
Y/E
p
Es Y E Y
b
Y E E E n.a. Y Y Y
pd
Y
v2
Y
pd
Y Y
INM-CM4 Y/E Y Y Y
b
Y
b
n.a. Y
fx
n.a. n.a. n.a. Y
so
n.a. n.a. Y
v0
n.a. Y Y
IPSL-CM5A-LR
18
Y/E
p
Y Y Y
e
Y
e
Y Y
e
Y
e
Y
e
n.a. Y n.a. Y
e
Y
v1
Y
e
Y Y
IPSL-CM5A-MR
18
Y/E
p
Y Y Y
e
Y
e
Y Y
e
Y
e
Y
e
n.a. Y n.a. Y
e
Y
v1
Y
e
Y Y
IPSL-CM5B-LR
18
Y
p
Y Y Y
e
Y
e
Y Y
e
Y
e
Y
e
n.a. Y n.a. Y
e
Y
v1
Y
e
Y Y
MIROC-ESM
19
Y/E
p
Y Y Y
f
Y
f
Y E E E n.a. Y
ic
Y
ic
Y
pd
Y
v3
Y
pd
Y Y
MIROC-ESM-CHEM
19
Y
p
Y Y E E Y E E E n.a. Y
ic
Y
ic
Y
pd
Y
v3
Y
pd
Y Y
MIROC4h
20
Y
p
Y Y Y
g
Y
g
Y E E E n.a. Y Y Y
pd
Y
v3
Y
pd
Y
cr
Y
MIROC5
20
Y
p
Y Y Y
f
Y
f
Y E E E n.a. Y
ic
Y
ic
Y
pd
Y
v3
Y
pd
Y
cr
Y
MPI-ESM-LR Y/E
p
Y Y Y
b
Y
b
Y Y
h
Y
h
Y
h
Y
h
n.a. n.a. Y
h
Y
v0
Y
h
Y Y
or
MPI-ESM-MR Y
p
Y Y Y
b
Y
b
Y Y
h
Y
h
Y
h
Y
h
n.a. n.a. Y
h
Y
v0
Y
h
Y Y
or
MPI-ESM-P Y
p
Y Y Y
b
Y
b
Y Y
h
Y
h
Y
h
Y
h
n.a. n.a. Y
h
Y
v0
Y
h
Y Y
or
MRI-CGCM3
21
Y Y Y Y
b
Y
b
Y E E E n.a. Y
ic
Y
ic
E
pd
E
v0
E
pd
Y Y
MRI-ESM1
22
E Y Y E E Es E E E n.a. Y
ic
Y
ic
E
pd
E
v0
E
pd
Y Y
NorESM1-M
23
Y
p
Y Y Y
a
Y
a
Y E E E n.a. Y Y E Y/E
st,v1
E
pd
Y Y
NorESM1-ME
23
Y/E
p
Y Y Y
a
Y
a
Y E E E n.a. Y Y E Y/E
st,v1
E
pd
Y Y
16
Jones et al. (2011)
17
Hardiman et al. (2012)
18
Dufresne et al. (2013)
19
Watanabe et al. (2011)
20
Komuro et al. (2012)
21
Yukimoto et al. (2012)
22
Adachi et al. (2013)
23
Iversen et al. (2013); Kirkevåg et al. (2013); Tjiputra et al. (2013)
Notes:
Model-specific references relating to forcing implementations:
1
Dix et al. (2013)
2
Wu et al. (2013); Xin et al. (2013a, 2013b)
3
Meehl et al. (2012); Gent et al. (2011)
4
Long et al. (2013); Meehl et al. (2012)
5
Meehl et al. (2013)
6
Calvo et al. (2012); Meehl et al. (2012)
7
Cagnazzo et al. (2013)
8
Voldoire et al. (2013)
9
Rotstayn et al. (2012)
10
Hazeleger et al. (2013)
11
Li et al. (2013c)
12
Bao et al. (2013)
13
Levy II et al. (2013)
14
Shindell et al. (2013a). GISS-E2-R and GISS-E2-H model variants are forced similarly and both
represented here as GISS-E2. Both -R and -H model versions have three variants: in physics version
1 (p1) aerosols and ozone are specified from pre-computed transient aerosol and ozone fields, in
physics version 2 (p2) aerosols and atmospheric chemistry are calculated online as a function of
atmospheric state and transient emissions inventories, while in physics version 3 (p3) atmospheric
composition is calculated as for p2 but the aerosol impacts on clouds (and hence the aerosol indirect
effect) is calculated interactively. In p1 and p2 variants the aerosol indirect effect is parameterized
following Hansen et al. (2005b).
15
HadGEM2-AO is forced in a similar way to HadGEM2-ES and HadGEM2-CC following Jones et al.
(2011), but tropospheric ozone, stratospheric ozone and land cover are prescribed.
Table 12.1 (continued)
(continued on next page)
1050
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Additional notes:
ce
Separate entries for CO
2
denote ‘concentration-driven’ and ‘emissions-driven’ experiments as indicated.
ac
‘Cloud albedo effect’ and ‘Cloud lifetime effect’ are classical terms (as used in AR4) to describe indirect effects of radiative
forcing associated with aerosols. They relate to the revised terminologies defined in Chapter 7 and used in AR5: ‘Radiative
forcing from aerosol–cloud interactions (RFaci)’ and ‘Effective radiative forcing from aerosol–cloud interactions (ERFaci)’.
RFaci equates to cloud albedo effect, while ERFaci is the effective forcing resulting from cloud albedo effect plus cloud
lifetime effect, including all rapid adjustments to cloud lifetime and thermodynamics (Section 7.1.3, Figure 7.3).
p
Physiological forcing effect of CO
2
via plant stomatal response and evapotranspiration (Betts et al., 2007) included.
rc
Separate entries denote different treatments used for radiation and chemistry respectively.
hf
Separate entries denote treatment for historical and future (RCPs) respectively.
a
Three-dimensional tropospheric ozone, stratospheric ozone, methane, and/or aerosol distributions specified as monthly
10-year mean concentrations, computed off-line using CAM-Chem – a modified version of CAM3.5 with interactive chemistry –
driven with specified emissions for the historical period (Lamarque et al., 2010) and RCPs (Lamarque et al., 2011) with sea
surface temperature and sea ice boundary conditions based on CCSM3’s projections for the closest corresponding AR4
scenarios.
b
Ozone prescribed using the original or slightly modified IGAC/SPARC ozone data set (Cionni et al., 2011); in some models this
data set is modified to add a future solar cycle and in some models tropospheric ozone is zonally averaged.
c
Linearized 2D ozone chemistry scheme (Cariolle and Teyssedre, 2007) including transport and photochemistry, reactive to
stratospheric chlorine concentrations but not tropospheric chemical emissions.
d
Ozone prescribed using the data set described in Hansen et al. (2007), with historical tropospheric ozone being calculated
by a CCM and stratospheric ozone taken from Randel and Wu (2007) in the past. Tropospheric ozone is held constant from
1990 onwards, while stratospheric ozone is constant from 1997 to 2003 and then returned linearly to its 1979 value over the
period 2004 to 2050.
e
For IPSL-CM5 model versions, ozone and aerosol concentrations are calculated semi-offline with the atmospheric general
circulation model including interactive chemistry and aerosol, following the four RCPs in the future (Dufresne et al., 2013;
Szopa et al., 2013). The same aerosol concentration fields (but not ozone) are also prescribed for the CNRM-CM5 model.
f
Ozone concentrations computed off-line by Kawase et al. (2011) using a CCM forced with CMIP5 emissions.
g
Ozone concentrations computed off-line by Sudo et al. (2003) for the historical period and Kawase et al. (2011) for the future.
h
Time dependent climatology based on simulations and observations; aerosols are distinguished only with respect to coarse
and fine mode, and anthropogenic and natural origins, not with respect to composition.
op
Separate entries denote different ozone chemistry precursors.
so
RFaci from sulphate aerosol only.
st
Separate entries denote stratosphere and troposphere respectively.
ic
Radiative effects of aerosols on ice clouds are represented.
pd
Prognostic or diagnostic scheme for dust/sea salt aerosol with emissions/concentrations determined by the model state
rather than externally prescribed.
fx
Fixed prescribed climatology of dust/sea salt aerosol concentrations with no year-to-year variability.
v0
Explosive volcanic aerosol returns rapidly in future to zero (or near-zero) background, like that in the pre-industrial control
experiment.
v1
Explosive volcanic aerosol returns rapidly in future to constant (average volcano) background, the same as in the pre-
industrial control experiment.
v2
Explosive volcanic aerosol returns slowly in future (over several decades) to constant (average volcano) background like that
in the pre-industrial control experiment.
v3
Explosive volcanic aerosol returns rapidly in future to near-zero background, below that in the pre-industrial control
experiment.
v4
Explosive volcanic aerosol set to zero in future, but constant (average volcano) background in the pre-industrial control
experiment.
v5
Explosive volcanic aerosol returns slowly in future (over several decades) to constant (average volcano) background, but zero
background in the pre-industrial control experiment.
cr
Land use change represented via crop change only.
or
Realistic time-varying orbital parameters for solar forcing (in historical period only for GISS-E2).
Table 12.1 (continued)
1051
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
very precisely. A repeated 11-year cycle for total solar irradiance (Lean
and Rind, 2009) is suggested for future projections but the periodicity
is not specified precisely as solar cycles vary in length. Some models
include the effect of orbital variations as well, but most do not. For
volcanic eruptions, no specific CMIP5 prescription is given for future
emissions or concentration data, the general recommendation being
that volcanic aerosols should either be omitted entirely both from the
control experiment and future projections or the same background
volcanic aerosols should be prescribed in both. This provides a con-
sistent framework for model intercomparison given a lack of knowl-
edge of when future large eruptions will occur. In general models have
adhered to this guidance, but there are variations in the background
volcanic aerosol levels chosen (zero or an average volcano back-
ground in general) and some cases, for example, Australian Commu-
nity Climate and Earth System Simulator (ACCESS)1.0 and ACCESS1.3
(Dix et al., 2013), where the background volcanic aerosol in future
differs significantly from that in the control experiment, with a small
effect on future RF.
For the other natural aerosols (dust, sea-salt, etc.), no emission or
concentration data are recommended. The emissions are potentially
computed interactively by the models themselves and may change
with climate, or prescribed from separate model simulations carried
out in the implementation of CMIP5 experiments, or simply held con-
stant. Natural aerosols (mineral dust and sea salt) are in a few cases
prescribed with no year-to-year variation (giving no transient forcing
effect), in some cases prescribed from data sets computed off-line as
described above, and in other cases calculated interactively via prog-
nostic or diagnostic calculations. The degree to which natural aerosol
emissions are interactive is effectively greater in some such models
than others, however, as mineral dust emissions are more constrained
when land vegetation cover is specified (e.g., as in Commonwealth
Scientific and Industrial Research Organisation (CSIRO)-Mk3.6.0) (Rot-
stayn et al., 2012) than when vegetation is allowed to evolve dynami-
cally (e.g., as in Hadley Centre new Global Environmental Model 2-ES
(HadGEM2-ES)) (Jones et al., 2011) (Table 9.A.1).
12.3.2.1 ‘Emissions-Driven’ versus ‘Concentration Driven’
Experiments
A novel feature within the CMIP5 experimental design is that experi-
ments with prescribed anthropogenic emissions are included in addi-
tion to classical experiments with prescribed concentration pathways
for WMGHGs (Taylor et al., 2012). The essential features of these two
classes of experiment are described in Box 6.4. The CMIP5 protocol
includes experiments in which ‘ESMs’ (models possessing at least a
carbon cycle, allowing for interactive calculation of atmospheric CO
2
or compatible emissions) and AOGCMs (that do not possess such an
interactive carbon cycle) are both forced with WMGHG concentration
pathways to derive a range of climate responses consistent with those
pathways from the two types of model. The range of climate responses
including climate–carbon cycle feedbacks can additionally be explored
in ESMs driven with emissions rather than concentrations, analogous
to Coupled Climate Carbon Cycle Model Intercomparison Project
(C
4
MIP) experiments (Friedlingstein et al., 2006)—see Box 6.4. Results
from the two types of experiment cannot be compared directly, but
they provide complementary information. Uncertainties in the forward
climate response driven with specified emissions or concentrations can
be derived from all participating models, while concentration-driven
ESM experiments also permit a policy-relevant diagnosis of the range
of anthropogenic carbon emissions compatible with the imposed con-
centration pathways (Hibbard et al., 2007; Moss et al., 2010).
WMGHG forcing implementations in CMIP5 concentration-driven
experiments conform closely in almost all cases to the standard proto-
col (Table 12.1; CO
2
, CH
4
, N
2
O, chlorofluorocarbons (CFCs)), imposing
an effective control over the RF due to WMGHGs across the multi-mod-
el ensemble, apart from the model spread arising from radiative trans-
fer codes (Collins et al., 2006b; Meehl et al., 2007b). The ability of ESMs
to determine their own WMGHG concentrations in emissions-driven
experiments means that RF due to WMGHGs is less tightly controlled
in such experiments. Even in concentration-driven experiments, many
models implement some emissions-driven forcing agents (more often
aerosols, but also ozone in some cases), leading to a potentially great-
er spread in both the concentrations and hence RF of those emis-
sions-driven agents.
12.3.2.2 Variations Between Model Forcing Implementations
Apart from the distinction between concentration-driven and emis-
sions-driven protocols, a number of variations are present in the imple-
mentation of forcing agents listed in Table 12.1, which generally arise
due to constraining characteristics of the model formulations, various
computational efficiency considerations or local implementation deci-
sions. In a number of models, off-line modelling using an aerosol chem-
istry climate model has been used to convert emissions into concentra-
tions compatible with the specific model formulation or characteristics.
As a result, although detailed prescriptions are given for the forcing
agents in CMIP5 experiments in emissions terms, individual modelling
approaches lead to considerable variations in their implementations
and consequential RFs. This was also the case in the ENSEMBLES mul-
ti-model projections, in which similar forcing agents to CMIP5 models
were applied but again with variations in the implementation of aer-
osol, ozone and land use forcings, prescribing the SRES A1B and E1
scenarios in a concentration-driven protocol (Johns et al., 2011) akin
to the CMIP5 protocol.
Methane, nitrous oxide and CFCs (typically with some aggregation of
the multiple gases) are generally prescribed in CMIP5 models as well-
mixed concentrations following the forcing data time series provid-
ed for the given scenarios. In a number of models (CESM1(WACCM),
GFDL-CM3, GISS-E2-p2, GISS-E2-p3, HadGEM2-ES and MRI-ESM1) the
three-dimensional concentrations in the atmosphere of some species
evolve interactively driven by the full emissions/sinks cycle (in some
cases constrained by prescribed concentrations at the surface, e.g.,
HadGEM2-ES for methane). In cases where the full emissions/sinks
cycle is modelled, the radiation scheme is usually passed the time-var-
ying 3-D concentrations, but some models prescribe different concen-
trations for the purpose of radiation.
Eyring et al. (2013) document, in greater detail than Table 12.1, the
implementations of tropospheric and stratospheric ozone in CMIP5
models, including their ozone chemistry schemes and modifications
applied to reference data sets in models driven by concentrations. In
1052
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
most models that prescribe ozone, concentrations are based on the
original or slightly modified CMIP5 standard ozone data set comput-
ed as part of the International Global Atmospheric Chemistry/Strat-
ospheric Processes and their Role in Climate (IGAC/SPARC) activity
(Cionni et al., 2011). In the stratosphere, this data set is based on
observations of the past (Randel and Wu, 2007) continued into the
future with the multi-model mean of 13 chemistry–climate models
(CCMs) projections following the SRES A1B (IPCC, 2000) and SRES
A1 adjusted halogen scenario (WMO, 2007). The stratospheric zonal
mean ozone field is merged with a 3-D tropospheric ozone time series
generated as the mean of two CCMs (Goddard Institute of Space
Studies-Physical Understanding of Composition-Climate Interactions
and Impacts (GISS-PUCCINI), Shindell et al., 2006; CAM3.5, Lamarque
et al., 2010) in the past and continued by one CCM (CAM3.5) in the
future. Some CMIP5 models (MIROC-ESM, MIROC4h, MIROC5 and
GISS-E2-p1) prescribe ozone concentrations using different data sets
but again following just one GHG scenario in the future for the projec-
tion of stratospheric ozone. In other models (e.g., Institut Pierre Simon
Laplace (IPSL)-CM5, CCSM4) ozone is again prescribed, but supplied as
concentrations from off-line computations using a related CCM. Some
models determine ozone interactively from specified emissions via
on-line atmospheric chemistry (CESM1(FASTCHEM), CESM1(WACCM),
CNRM-CM5, GFDL-CM3, GISS-E2-p2, GISS-E2-p3, MIROC-ESM-CHEM,
MRI-ESM1; and HadGEM2-ES for tropospheric ozone only). Computing
ozone concentrations interactively allows the fast coupling between
chemistry and climate to be captured, but modelling of chemistry pro-
cesses is sometimes simplified (CNRM-CM5, CESM(FASTCHEM)) in
comparison with full complexity CCMs to reduce the computational
cost. Compared to CMIP3, in which all models prescribed ozone and
around half of them used a fixed ozone climatology, this leads to sub-
stantial improvement to ozone forcings in CMIP5, although differences
remain among the models with interactive chemistry.
For atmospheric aerosols, either aerosol precursor emissions-driven
or concentration-driven forcings are applied depending on individu-
al model characteristics (see Sections 7.3 and 7.4 for an assessment
of aerosols processes including aerosol–radiation and aerosol–cloud
interactions). A larger fraction of models in CMIP5 than CMIP3 pre-
scribe aerosol precursor emissions rather than concentrations. Many
still prescribe concentrations pre-computed either using a directly relat-
ed aerosol CCM or from output of another, complex, emissions-driven
aerosol chemistry model within the CMIP5 process. As for ozone, aer-
osol concentrations provided from off-line simulations help to reduce
the computational burden of the projections themselves. For several
of the concentration-driven models (CCSM4, IPSL-CM5A variants,
MPI-ESM-LR, MPI-ESM-MR), additional emissions-driven simulations
have been undertaken to tailor the prescribed concentrations closely
to the model’s individual aerosol–climate characteristics. Lamarque et
al. (2010, 2011) provided the recommended CMIP5 aerosols data set
which has been used in several of the models driven by concentrations.
Compared with the CMIP3 models, a much larger fraction of CMIP5
models now incorporate black and organic carbon aerosol forcings.
Also, a larger fraction of CMIP5 than CMIP3 models now includes a
range of processes that combine in the effective RF from aerosol–
cloud interactions (ERFaci; see Section 7.1.3 and Figure 7.3). Previ-
ously such processes were generally termed aerosol indirect effects,
usually separated into cloud albedo (or first indirect) effect and cloud
lifetime (or second indirect) effect. Many CMIP5 models only include
the interaction between sulphate aerosol and cloud, and the majority
of them only model the effect of aerosols on cloud albedo rather than
cloud lifetime (Table 12.1). No CMIP5 models represent urban aero-
sol pollution explicitly so that is not listed in Table 12.1 (see Section
11.3.5.2 for discussion of future air quality). Only one model (GISS-E2)
explicitly includes nitrate aerosol as a separate forcing, though it is
also included within the total aerosol mixture in the Max Planck Insti-
tute-Earth System Model (MPI-ESM) model versions.
Land use change is typically applied by blending anthropogenic land
surface disturbance via crop and pasture fraction changes with under-
lying land cover maps of natural vegetation, but model variations
in the underlying land cover maps and biome modelling mean that
the land use forcing agent is impossible to impose in a completely
common way at present (Pitman et al., 2009). Most CMIP5 models rep-
resent crop and pasture disturbance separately, while some (Canadian
Earth System Model (CanESM2), MIROC4h, MIROC5) represent crop
but not pasture. Some models (e.g., HadGEM2-ES, MIROC-ESM and
MPI-ESM versions) allow a dynamical representation of natural vege-
tation changes alongside anthropogenic disturbance (see also Sections
9.4.4.3 and 9.4.4.4).
Treatment of the CO
2
emissions associated with land cover chang-
es is also model dependent. Some models do not account for land
cover changes at all, some simulate the biophysical effects but are
still forced externally by land cover change induced CO
2
emissions (in
emissions-driven simulations), while the most advanced ESMs simu-
late both biophysical effects of land cover changes and their associ-
ated CO
2
emissions.
12.3.3 Synthesis of Projected Global Mean Radiative
Forcing for the 21st Century
Quantification of future global mean RF is of interest as it is directly
related to changes in the global energy balance of the climate system
and resultant climate change. Chapter 8 discusses RF concepts and
methods for computing it that form the basis of analysis directly from
the output of model projections.
We assess three related estimates of projected global mean forc-
ing and its range through the 21st century in the context of forcing
estimated for the recent past (Figure 12.4). The estimates used are:
the total forcings for the defined RCP scenarios, harmonized to RF in
the past (Meinshausen et al., 2011a; Meinshausen et al., 2011c); the
total effective radiative forcing (ERF) estimated from CMIP5 models
through the 21st century for the four RCP experiments (Forster et al.,
2013); and that estimated from models in the Atmospheric Chemistry
and Climate Model Intercomparison Project (ACCMIP; Lamarque et
al., 2013—see Section 8.2.2 ) for RCP time-slice experiments (Shindell
et al., 2013b). Methodological differences mean that whereas CMIP5
estimates include both natural and anthropogenic forcings based
entirely on ERF, ACCMIP estimates anthropogenic composition forcing
only (neglecting forcing changes due to natural, i.e., solar and volca-
nic, and land use factors) based on a combination of ERF for aerosols
and RF for WMGHG (see Section 8.5.3). Note also that total forcing
for the defined RCP scenarios is based on Meinshausen et al. (2011c)
1053
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
1980 2000 2020 2040 2060 2080 2100
-2
0
2
4
6
8
historical
historical (CMIP5)
RCP2.6
RCP2.6 (CMIP5)
RCP4.5
RCP4.5 (CMIP5)
RCP6.0
RCP6.0 (CMIP5)
RCP8.5
RCP8.5 (CMIP5)
historical (ACCMIP)
RCP2.6 (ACCMIP)
RCP4.5 (ACCMIP)
RCP6.0 (ACCMIP)
RCP8.5 (ACCMIP)
22
17
19
14
19
20C3M+A1B (CMIP3)
+/- 1 Stdev
23
21
but combining total anthropogenic ERF (allowing for efficacies of the
various anthropogenic forcings as in Figure 12.3) with natural (solar
and volcanic) RF.
The CMIP5 multi-model ensemble mean ERF at 2100 (relative to an
1850–1869 base period) is 2.2, 3.8, 4.8 and 7.6 W m
–2
respectively for
RCP2.6, RCP4.5, RCP6.0 and RCP8.5 concentration-driven projections,
with a 1-σ range based on annual mean data for year 2100 of about
±0.5 to 1.0 W m
–2
depending on scenario (lowest for RCP2.6 and high-
est for RCP8.5). The CMIP5-based ERF estimates are close to the total
forcing at 2100 (relative to an 1850–1859 base period) of 2.4, 4.0, 5.2
and 8.0 W m
–2
as defined for the four RCPs.
The spread in ERF indicated from CMIP5 model results with specified
GHG concentration pathways is broadly consistent with that found for
Figure 12.4 | Global mean radiative forcing (RF, W m
–2
) between 1980 and 2100 estimated by alternative methods. The baseline is circa 1850 but dependent on the methods.
Dashed lines indicate the total anthropogenic plus natural (solar and volcanic) RF for the RCP scenarios as defined by Meinshausen et al. (2011c), taking into account the efficacies
of the various anthropogenic forcings (Meinshausen et al., 2011a), normalized by the mean between 1850 and 1859. Solid lines are multi-model mean effective radiative forcing
(ERF) realized in a subset of CMIP5 models for the concentration-driven historical experiment and RCP scenarios, normalized either with respect to the 1850–1869 base period
or with respect to the pre-industrial control simulation (Forster et al., 2013). (The subset of CMIP5 models included is defined by Table 1 of Forster et al. (2013) but omitting the
FGOALS-s2 (Flexible Global Ocean-Atmosphere-Land System) model, the historical and RCP simulations of which were subsequently withdrawn from the CMIP5 archive.) This
CMIP5-based estimate assumes each model has an invariant climate feedback parameter, calculated from abrupt 4 × CO
2
experiments using the method of Gregory et al. (2004).
Each individual CMIP5 model’s forcing estimate is an average over all available ensemble members, and a 1-σ inter-model range around the multi-model mean is shaded in light
colour. Grey or coloured vertical bars illustrate the 1-σ range (68% confidence interval) of anthropogenic composition forcing (excluding natural and land use change forcings,
based on ERF for aerosols combined with RF for WMGHG) estimated in ACCMIP models (Shindell et al., 2013b) for time slice experiments at 1980, 2000, 2030 (RCP8.5 only) and
2100 (all RCPs). The ACCMIP ranges plotted have been converted from the 5 to 95% ranges given in Shindell et al. (2013b) (Table 8) to a 1-σ range. Note that the ACCMIP bars at
1980 and 2100 are shifted slightly to aid clarity. The mean ERF diagnosed from 21 CMIP3 models for the SRES A1B scenario, as in Forster and Taylor (2006), is also shown (thick
green line) with a 1-σ range (thinner green lines). The number of models included in CMIP3 and CMIP5 ensemble means is shown colour coded. (See Tables AII.6.8 to AII.6.10.
Note that the CMIP5 model ranges given in Table AII.6.10 are based on decadal averages and therefore differ slightly from the ranges based on annual data shown in this figure.)
CMIP3 models for the A1B scenario using the corresponding method
(Forster and Taylor, 2006). As for CMIP3 models, part of the forcing
spread in CMIP5 models (Forster et al., 2013) is consistent with differ-
ences in GHG forcings arising from the radiative transfer codes (Col-
lins et al., 2006b). Aerosol forcing implementations in CMIP5 models
also vary considerably, however (Section 12.3.2), leading to a spread
in aerosol concentrations and forcings which contributes to the overall
model spread. A further small source of spread in CMIP5 results pos-
sibly arises from an underlying ambiguity in the CMIP5 experimental
design regarding the volcanic forcing offset between the historical
experiment versus the pre-industrial control experiment. Most models
implement zero volcanic forcing in the control experiment but some
use constant negative forcing equal to the time-mean of historical
volcanic forcing (see Table 12.1 and Section 12.3.2). The effect of this
volcanic forcing offset persists into the future projections.
1054
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
ACCMIP projected forcing at 2030 (for RCP8.5) and 2100 (all RCPs) is
systematically higher than corresponding CMIP5 ERF, although with
some overlap between 1-σ ranges. CMIP5 and ACCMIP comprise dif-
ferent sets of models and they are related in many but not all cases
(Section 8.2.2). Confining analysis to a subset of closely related models
also gives higher forcing estimates from ACCMIP compared to CMIP5
so the discrepancy in multi-model ensemble mean forcings appears
unrelated to the different model samples associated with the two
methods of estimation. The discrepancy is thought to originate mostly
from differences in the underlying methodologies used to estimate RF,
but is not yet well understood (see also Section 8.5.3).
There is high confidence in projections from ACCMIP models (Shindell
et al., 2013b) based on the GISS-E2 CMIP5 simulations (Shindell et al.,
2013a) and an earlier study with a version of the HadGEM2-ES model
related to that used in CMIP5 (Bellouin et al., 2011), consistent with
understanding of the processes controlling nitrate formation (Adams
et al., 2001), that nitrate aerosols (which provide a negative forcing)
will increase substantially over the 21st century under the RCPs (Sec-
tion 8.5.3, Figure 8.20). The magnitude of total aerosol-related forcing
(also negative in sign) will therefore tend to be underestimated in the
CMIP5 multi-model mean ERF, as nitrate aerosol has been omitted as a
forcing from almost all CMIP5 models.
Natural RF variations are, by their nature, difficult to project reliably
(see Section 8.4). There is very high confidence that Industrial Era nat-
ural forcing has been a small fraction of the (positive) anthropogenic
forcing except for brief periods following large volcanic eruptions (Sec-
tions 8.5.1 and 8.5.2). Based on that assessment and the assumption
that variability in natural forcing remains of a similar magnitude and
character to that over the Industrial Era, total anthropogenic forcing
relative to pre-industrial, for any of the RCP scenarios through the 21st
century, is very likely to be greater in magnitude than changes in natu-
ral (solar plus volcanic) forcing on decadal time scales.
In summary, global mean forcing projections derived from climate
models exhibit a substantial range for the given RCP scenarios in con-
centration-driven experiments, contributing to the projected global
mean temperature range (Section 12.4.1). Forcings derived from
ACCMIP models for 2100 are systematically higher than those estimat-
ed from CMIP5 models for reasons that are not fully understood but
are partly due to methodological differences. The multi-model mean
estimate of combined anthropogenic plus natural forcing from CMIP5
is consistent with indicative RCP forcing values at 2100 to within 0.2
to 0.4 W m
–2
.
12.4 Projected Climate Change over the
21st Century
12.4.1 Time-Evolving Global Quantities
12.4.1.1 Projected Changes in Global Mean Temperature and
Precipitation
A consistent and robust feature across climate models is a continua-
tion of global warming in the 21st century for all the RCP scenarios
(Figure 12.5 showing changes in concentration-driven model simu-
lations). Temperature increases are almost the same for all the RCP
scenarios during the first two decades after 2005 (see Figure 11.25).
At longer time scales, the warming rate begins to depend more on
the specified GHG concentration pathway, being highest (>0.3°C per
decade) in the highest RCP8.5 and significantly lower in RCP2.6, par-
ticularly after about 2050 when global surface temperature response
stabilizes (and declines thereafter). The dependence of global temper-
ature rise on GHG forcing at longer time scales has been confirmed by
several studies (Meehl et al., 2007b). In the CMIP5 ensemble mean,
global warming under RCP2.6 stays below 2°C above 1850-1900
levels throughout the 21st century, clearly demonstrating the potential
of mitigation policies (note that to translate the anomalies in Figure
12.5 into anomalies with respect to that period, an assumed 0.61°C
of observed warming since 1850–1900, as discussed in Section 2.4.3,
should be added). This is in agreement with previous studies of aggres-
sive mitigation scenarios (Johns et al., 2011; Meehl et al., 2012). Note,
however, that some individual ensemble members do show warming
exceeding 2°C above 1850-1900 (see Table 12.3). As for the other
pathways, global warming exceeds 2°C within the 21st century under
RCP4.5, RCP6.0 and RCP8.5, in qualitative agreement with previous
studies using the SRES A1B and A2 scenarios (Joshi et al., 2011). Global
mean temperature increase exceeds 4°C under RCP8.5 by 2100. The
CMIP5 concentration-driven global temperature projections are broad-
ly similar to CMIP3 SRES scenarios discussed in AR4 (Meehl et al.,
2007b) and Section 12.4.9, although the overall range of the former
is larger primarily because of the low-emission mitigation pathway
RCP2.6 (Knutti and Sedláček, 2013).
The multi-model global mean temperature changes under different
RCPs are summarized in Table 12.2. The relationship between cumu-
lative anthropogenic carbon emissions and global temperature is
assessed in Section 12.5 and only concentration-driven models are
42 models
39
25
42
32
12
17
12
Figure 12.5 | Time series of global annual mean surface air temperature anomalies
(relative to 1986–2005) from CMIP5 concentration-driven experiments. Projections are
shown for each RCP for the multi-model mean (solid lines) and the 5 to 95% range
(±1.64 standard deviation) across the distribution of individual models (shading). Dis-
continuities at 2100 are due to different numbers of models performing the exten-
sion runs beyond the 21st century and have no physical meaning. Only one ensemble
member is used from each model and numbers in the figure indicate the number of
different models contributing to the different time periods. No ranges are given for the
RCP6.0 projections beyond 2100 as only two models are available.
1055
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
RCP2.6 (ΔT in °C) RCP4.5 (ΔT in °C) RCP6.0 (ΔT in °C) RCP8.5 (ΔT in °C)
Global: 2046–2065 1.0 ± 0.3 (0.4, 1.6) 1.4 ± 0.3 (0.9, 2.0) 1.3 ± 0.3 (0.8, 1.8) 2.0 ± 0.4 (1.4, 2.6)
2081–2100 1.0 ± 0.4 (0.3, 1.7) 1.8 ± 0.5 (1.1, 2.6) 2.2 ± 0.5 (1.4, 3.1) 3.7 ± 0.7 (2.6, 4.8)
2181–2200 0.7 ± 0.4 (0.1, 1.3) 2.3 ± 0.5 (1.4, 3.1) 3.7 ± 0.7 (-,-) 6.5 ± 2.0 (3.3, 9.8)
2281–2300 0.6 ± 0.3 (0.0, 1.2) 2.5 ± 0.6 (1.5, 3.5) 4.2 ± 1.0 (-,-) 7.8 ± 2.9 (3.0, 12.6)
Land: 2081–2100 1.2 ± 0.6 (0.3, 2.2) 2.4 ± 0.6 (1.3, 3.4) 3.0 ± 0.7 (1.8, 4.1) 4.8 ± 0.9 (3.4, 6.2)
Ocean: 2081–2100 0.8 ± 0.4 (0.2, 1.4) 1.5 ± 0.4 (0.9, 2.2) 1.9 ± 0.4 (1.1, 2.6) 3.1 ± 0.6 (2.1, 4.0)
Tropics: 2081–2100 0.9 ± 0.3 (0.3, 1.4) 1.6 ± 0.4 (0.9, 2.3) 2.0 ± 0.4 (1.3, 2.7) 3.3 ± 0.6 (2.2, 4.4)
Polar: Arctic: 2081–2100 2.2 ± 1.7 (-0.5, 5.0) 4.2 ± 1.6 (1.6, 6.9) 5.2 ± 1.9 (2.1, 8.3) 8.3 ± 1.9 (5.2, 11.4)
Polar: Antarctic: 2081–2100 0.8 ± 0.6 (-0.2, 1.8) 1.5 ± 0.7 (0.3, 2.7) 1.7 ± 0.9 (0.2, 3.2) 3.1 ± 1.2 (1.1, 5.1)
included here. Warming in 2046–2065 is slightly larger under RCP4.5
compared to RCP6.0, consistent with its greater total anthropogenic
forcing at that time (see Table A.II.6.12). For all other periods the mag-
nitude of global temperature change increases from RCP2.6 to RCP8.5.
Beyond 2100, RCP2.6 shows a decreasing trend whereas under all
other RCPs warming continues to increase. Also shown in Table 12.2
are projected changes at 2081–2100 averaged over land and ocean
separately as well as area-weighted averages over the Tropics (30°S
to 30°N), Arctic (67.5°N to 90°N) and Antarctic (90°S to 55°S) regions.
Surface air temperatures over land warm more than over the ocean,
and northern polar regions warm more than the tropics. The excess of
land mass in the Northern Hemisphere (NH) in comparison with the
Southern Hemisphere (SH), coupled with the greater uptake of heat by
the Southern Ocean in comparison with northern ocean basins means
that the NH generally warms more than the SH. Arctic warming is much
greater than in the Antarctic, due to the presence of the Antarctic ice
sheet and differences in local responses in snow and ice. Mechanisms
behind these features of warming are discussed in Section 12.4.3.
Maps and time series of regional temperature changes are displayed in
Annex I and regional averages are discussed in Section 14.8.1.
Global annual multi-model mean temperature changes above 1850-
1900 are listed in Table 12.3 for the 2081–2100 period (assuming
0.61°C warming since 1850–1900 as discussed in Section 2.4.3)
along with the percentage of 2081–2100 projections from the CMIP5
models exceeding policy-relevant temperature levels under each RCP.
These complement a similar discussion for the near-term projections
in Table 11.3 which are based on the CMIP5 ensemble as well as
evidence (discussed in Sections 10.3.1, 11.3.2.1.1 and 11.3.6.3) that
some CMIP5 models have a higher sensitivity to GHGs and a larger
response to other anthropogenic forcings (dominated by the effects
of aerosols) than the real world (medium confidence). The percent-
age calculations for the long-term projections in Table 12.3 are based
solely on the CMIP5 ensemble, using one ensemble member for each
model. For these long-term projections, the 5 to 95% ranges of the
CMIP5 model ensemble are considered the likely range, an assess-
ment based on the fact that the 5 to 95% range of CMIP5 models’
TCR coincides with the assessed likely range of the TCR (see Section
12.4.1.2 below and Box 12.2). Based on this assessment, global mean
temperatures averaged in the period 2081–2100 are projected to
likely exceed 1.5°C above 1850-1900 for RCP4.5, RCP6.0 and RCP8.5
(high confidence). They are also likely to exceed 2°C above 1850-1900
for RCP6.0 and RCP8.5 (high confidence) and more likely than not
to exceed 2°C for RCP4.5 (medium confidence). Temperature change
above 2°C under RCP2.6 is unlikely but is assessed only with medium
confidence as some CMIP5 ensemble members do produce a global
mean temperature change above 2°C. Warming above 4°C by 2081–
2100 is unlikely in all RCPs (high confidence) except RCP8.5. Under
the latter, the 4°C global temperature level is exceeded in more than
half of ensemble members, and is assessed to be about as likely as not
(medium confidence). Note that the likelihoods of exceeding specific
temperature levels show some sensitivity to the choice of reference
period (see Section 11.3.6.3).
CMIP5 models on average project a gradual increase in global precip-
itation over the 21st century: change exceeds 0.05 mm day
–1
(~2%
of global precipitation) and 0.15 mm day
–1
(~5% of global precipi-
tation) by 2100 in RCP2.6 and RCP8.5, respectively. The relationship
between global precipitation and global temperature is approximately
linear (Figure 12.6). The precipitation sensitivity, that is, the change of
global precipitation with temperature, is about 1 to 3% °C
–1
in most
models, tending to be highest for RCP2.6 and RCP4.5 (Figure 12.7;
note that only global values are discussed in this section, ocean and
land changes are discussed in Section 12.4.5.2). These behaviours are
consistent with previous studies, including CMIP3 model projections
for SRES scenarios and AR4 constant composition commitment exper-
iments (Meehl et al., 2007b), and ENSEMBLES multi-model results for
SRES A1B and E1 scenarios (Johns et al., 2011).
The processes that govern global precipitation changes are now well
understood and have been presented in Section 7.6. They are briefly
summarized here and used to interpret the long-term projected chang-
es. The precipitation sensitivity (about 1 to 3% °C
–1
) is very different
from the water vapour sensitivity (~7% °C
–1
) as the main physical
Table 12.2 | CMIP5 annual mean surface air temperature anomalies (°C) from the 1986–2005 reference period for selected time periods, regions and RCPs. The multi-model
mean ±1 standard deviation ranges across the individual models are listed and the 5 to 95% ranges from the models’ distribution (based on a Gaussian assumption and obtained
by multiplying the CMIP5 ensemble standard deviation by 1.64) are given in brackets. Only one ensemble member is used from each model and the number of models differs for
each RCP (see Figure 12.5) and becomes significantly smaller after 2100. No ranges are given for the RCP6.0 projections beyond 2100 as only two models are available. Using
Hadley Centre/Climate Research Unit gridded surface temperature data set 4 (HadCRUT4) and its uncertainty estimate (5 to 95% confidence interval), the observed warming to the
1986–2005 reference period (see Section 2.4.3) is 0.61°C ± 0.06°C (1850–1900), 0.30°C ± 0.03°C (1961–1990), 0.11°C ± 0.02°C (1980–1999). Decadal values are provided
in Table AII.7.5, but note that percentiles of the CMIP5 distributions cannot directly be interpreted in terms of calibrated language.
1056
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Table 12.3 | CMIP5 global annual mean temperature changes above 1850-1900 for the 2081–2100 period of each RCP scenario (mean, ±1 standard deviation and 5 to 95%
ranges based on a Gaussian assumption and obtained by multiplying the CMIP5 ensemble standard deviation by 1.64), assuming 0.61°C warming has occurred prior to 1986–2005
(second column). For a number of temperature levels (1°C, 1.5°C, 2°C, 3°C and 4°C), the proportion of CMIP5 model projections for 2081–2100 above those levels under each
RCP scenario are listed. Only one ensemble member is used for each model.
laws that drive these changes also differ. Water vapour increases are
primarily a consequence of the Clausius–Clapeyron relationship asso-
ciated with increasing temperatures in the lower troposphere (where
most atmospheric water vapour resides). In contrast, future precipi-
tation changes are primarily the result of changes in the energy bal-
ance of the atmosphere and the way that these later interact with
Figure 12.6 | Global mean precipitation (mm day
–1
) versus temperature (°C) changes
relative to 1986–2005 baseline period in CMIP5 model concentrations-driven projec-
tions for the four RCPs for (a) means over decadal periods starting in 2006 and over-
lapped by 5 years (2006–2015, 2011–2020, up to 2091–2100), each line representing
a different model (one ensemble member per model) and (b) corresponding multi-model
means for each RCP.
0 1 2 3 4 5 6
Temperature change (°C)
0.0
0.1
0.2
0.3
Precipitation change (mm day
-1
)
RCP2.6
RCP4.5
RCP6.0
RCP8.5
a
0 1 2 3 4 5 6
Temperature change (°C)
0.0
0.1
0.2
0.3
Precipitation change (mm day
-1
)
RCP2.6
RCP4.5
RCP6.0
RCP8.5
b
T (°C)
2081–2100
T > +1.0°C T > +1.5°C T > +2.0°C T > +3.0°C T > +4.0°C
RCP2.6 1.6 ± 0.4 (0.9, 2.3) 94% 56% 22% 0% 0%
RCP4.5 2.4 ± 0.5 (1.7, 3.2) 100% 100% 79% 12% 0%
RCP6.0 2.8 ± 0.5 (2.0, 3.7) 100% 100% 100% 36% 0%
RCP8.5 4.3 ± 0.7 (3.2, 5.4) 100% 100% 100% 100% 62%
circulation, moisture and temperature (Mitchell et al., 1987; Boer, 1993;
Vecchi and Soden, 2007; Previdi, 2010; O’Gorman et al., 2012). Indeed,
the radiative cooling of the atmosphere is balanced by latent heat-
ing (associated with precipitation) and sensible heating. Since AR4,
the changes in heat balance and their effects on precipitation have
been analyzed in detail for a large variety of forcings, simulations and
models (Takahashi, 2009a; Andrews et al., 2010; Bala et al., 2010; Ming
et al., 2010; O’Gorman et al., 2012; Bony et al., 2013).
An increase of CO
2
decreases the radiative cooling of the troposphere
and reduces precipitation (Andrews et al., 2010; Bala et al., 2010). On
longer time scales than the fast hydrological adjustment time scale
(Andrews et al., 2010; Bala et al., 2010; Cao et al., 2012; Bony et al.,
2013), the increase of CO
2
induces a slow increase of temperature and
water vapour, thereby enhancing the radiative cooling of the atmos-
phere and increasing global precipitation (Allen and Ingram, 2002;
Yang et al., 2003; Held and Soden, 2006). Even after the CO
2
forcing
stabilizes or begins to decrease, the ocean continues to warm, which
then drives up global temperature, evaporation and precipitation. In
addition, nonlinear effects also affect precipitation changes (Good et
al., 2012). These different effects explain the steepening of the precip-
itation versus temperature relationship in RCP2.6 and RCP4.5 scenari-
os (Figure 12.6), as RF stabilizes and/or declines from the mid-century
(Figure 12.4). In idealized CO
2
ramp-up/ramp-down experiments, this
effect produces an hydrological response overshoot (Wu et al., 2010).
An increase of absorbing aerosols warms the atmosphere and reduces
precipitation, and the surface temperature response may be too small
to compensate this decrease (Andrews et al., 2010; Ming et al., 2010;
Shiogama et al., 2010a). Change in scattering aerosols or incoming
solar radiation modifies global precipitation mainly via the response of
the surface temperature (Andrews et al., 2009; Bala et al., 2010).
The main reasons for the inter-model spread of the precipitation sen-
sitivity estimate among GCMs have not been fully understood. Never-
theless, spread in the changes of the cloud radiative effect has been
shown to have an impact (Previdi, 2010), although the effect is less
important for precipitation than it is for the climate sensitivity esti-
mate (Lambert and Webb, 2008). The lapse rate plus water vapour
feedback and the response of the surface heat flux (Previdi, 2010;
O’Gorman et al., 2012), the shortwave absorption by water vapour
(Takahashi, 2009b) or by aerosols, have been also identified as impor-
tant factors.
Global precipitation sensitivity estimates from observations are
very sensitive to the data and the time period considered. Some
1057
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Figure 12.7 | Percentage changes over the 21st century in global, land and ocean pre-
cipitation per degree Celsius of global warming in CMIP5 model concentration-driven
projections for the four RCP scenarios. Annual mean changes are calculated for each
year between 2006 and 2100 from one ensemble member per model relative to its
mean precipitation and temperature for the 1986–2005 baseline period, and the gradi-
ent of a least-squares fit through the annual data is derived. Land and ocean derived
values use global mean temperature in the denominator of dP/dT. Each coloured
symbol represents a different model, the same symbol being used for the same model
for different RCPs and larger black squares being the multi-model mean. Also shown
for comparison are global mean results for ENSEMBLES model concentrations-driven
projections for the E1 and A1B scenarios (Johns et al., 2011), in this case using a least-
squares fit derived over the period 2000–2099 and taking percentage changes relative
to the 1980–1999 baseline period. Changes of precipitation over land and ocean are
discussed in Section 12.4.5.2.
observational studies suggest precipitation sensitivity values higher
than model estimates (Wentz et al., 2007; Zhang et al., 2007), although
more recent studies suggest consistent values (Adler et al., 2008; Li et
al., 2011b).
12.4.1.2 Uncertainties in Global Quantities
Uncertainties in global mean quantities arise from variations in internal
natural variability, model response and forcing pathways. Table 12.2
gives two measures of uncertainty in the CMIP5 model projections,
the standard deviation and the 5 to 95% range across the ensemble’s
distribution. Because CMIP5 was not designed to explore fully the
uncertainty range in projections (see Section 12.2), neither its stand-
ard deviation nor its range can be interpreted directly as an uncer-
tainty statement about the corresponding real quantities, and other
techniques and arguments to assess uncertainty in future projections
must be considered. Figure 12.8 summarizes the uncertainty ranges
in global mean temperature changes at the end of the 21st century
under the various scenarios quantified by various methods. Individual
CMIP5 models are shown by red crosses. Red bars indicate mean and
5 to 95% percentiles based on assuming a normal distribution for the
CMIP5 sample (i.e., ±1.64 standard deviations). Estimates from the
simple climate carbon cycle Model for the Assessment of Greenhouse
Gas-Induced Climate Change (MAGICC; Meinshausen et al., 2011a;
Meinshausen et al., 2011b) calibrated to C
4
MIP (Friedlingstein et al.,
2006) carbon cycle models, assuming a PDF for climate sensitivity that
corresponds to the assessment of IPCC AR4 (Meehl et al., 2007b, Box
10.2), are given as yellow bars (Rogelj et al., 2012). Note that not all
2.6 4.5 6.0 8.5 2.6 4.5 6.0 8.5 2.6 4.5 6.0 8.5 E1 A1B
-4
-2
0
2
4
6
dP/dT (% °C
-1
)
CMIP5
Global
CMIP5
Land
CMIP5
Ocean
ENSEMBLES
Global
RCPRCP RCP
models have simulated all scenarios. To test the effect of undersam-
pling, and to generate a consistent set of uncertainties across scenarios,
a step response method that estimates the total warming as sum of
responses to small forcing steps (Good et al., 2011a) is used to emulate
23 CMIP5 models under the different scenarios (those 23 models that
supplied the necessary simulations to compute the emulators, i.e., CO
2
step change experiments). This provides means and ranges (5 to 95%)
that are comparable across scenarios (blue). See also Section 12.4.9 for
a discussion focussed on the differences between CMIP3 and CMIP5
projections of global average temperature changes.
For the CO
2
concentration-driven simulations (Figure 12.8a), the dom-
inant driver of uncertainty in projections of global temperature for the
higher RCPs beyond 2050 is the transient climate response (TCR), for
RCP2.6, which is closer to equilibrium by the end of the century, it is
both the TCR and the equilibrium climate sensitivity (ECS). In a tran-
sient situation, the ratio of temperature to forcing is approximately
constant and scenario independent (Meehl et al., 2007b, Appendix
10.A.1; Gregory and Forster, 2008; Knutti et al., 2008b; Good et al.,
2013). Therefore, the uncertainty in TCR maps directly into the uncer-
tainty in global temperature projections for the RCPs other than
RCP2.6. The assessed likely range of TCR based on various lines of
evidence (see Box 12.2) is similar to the 5 to 95% percentile range
of TCR in CMIP5. In addition, the assessed likely range of ECS is also
consistent with the CMIP5 range (see Box 12.2). There is little evidence
that the CMIP5 models are significantly over- or underestimating the
RF. The RF uncertainty is small compared to response uncertainty (see
Figure 12.4), and is considered by treating the 5 to 95% as a likely
rather than very likely range. Kuhlbrodt and Gregory (2012) suggest
that models might be overestimating ocean heat uptake, as previously
suggested by Forest et al. (2006), but observationally constrained esti-
mates of TCR are unaffected by that. The ocean heat uptake efficiency
does not contribute much to the spread of TCR (Knutti and Tomassini,
2008; Kuhlbrodt and Gregory, 2012).
Therefore, for global mean temperature projections only, the 5 to 95%
range (estimated as 1.64 times the sample standard deviation) of the
CMIP5 projections can also be interpreted as a likely range for future
temperature change between about 2050 and 2100. Confidence in this
assessment is high for the end of the century because the warming
then is dominated by CO
2
and the TCR. Confidence is only medium for
mid-century when the contributions of RF and initial conditions to the
total temperature response uncertainty are larger. The likely ranges are
an expert assessment, taking into account many lines of evidence, in
much the same way as in AR4 (Figure SPM.5), and are not probabilistic.
The likely ranges for 2046–2065 do not take into account the possible
influence of factors that lead to near-term (2016–2035) projections of
global mean surface temperature (GMST) that are somewhat cooler
than the 5 to 95% model ranges (see Section 11.3.6), because the
influence of these factors on longer term projections cannot be quan-
tified. A few recent studies indicate that some of the models with the
strongest transient climate response might overestimate the near term
warming (Otto et al., 2013; Stott et al., 2013) (see Sections 10.8.1,
11.3.2.1.1), but there is little evidence of whether and how much that
affects the long-term warming response. One perturbed physics ensem-
ble combined with observations indicates warming that exceeds the
AR4 at the top end but used a relatively short time period of warming
1058
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
(50 years) to constrain the models’ projections (Rowlands et al., 2012)
(see Sections 11.3.2.1.1 and 11.3.6.3). GMSTs for 2081–2100 (rela-
tive to 1986–2005) for the CO
2
concentration driven RCPs is therefore
assessed to likely fall in the range 0.3°C to 1.7°C (RCP2.6), 1.1°C to
2.6°C (RCP4.5), 1.4°C to 3.1°C (RCP6.0), and 2.6°C to 4.8°C (RCP8.5)
estimated from CMIP5. Beyond 2100, the number of CMIP5 simula-
tions is insufficient to estimate a likely range. Uncertainties before
2050 are assessed in Section 11.3.2.1.1. The assessed likely range is
very similar to the range estimated by the pulse response model, sug-
gesting that the different sample of models for the different RCPs are
not strongly affecting the result, and providing further support that
this pulse response technique can be used to emulate temperature and
ocean heat uptake in Chapter 13 and Section 12.4.9. The results are
consistent with the probabilistic results from MAGICC, which for the
lower RCPs have a slightly narrower range due to the lack of inter-
nal variability in the simple model, and the fact that non-CO
2
forcings
are treated more homogeneously than in CMIP5 (Meinshausen et al.,
2011a, 2011b). This is particularly pronounced for RCP2.6 where the
CMIP5 range is substantially larger, partly due to the larger fraction of
non-CO
2
forcings in that scenario.
The uncertainty estimate in AR4 for the SRES scenarios was –40% to
+60% around the CMIP3 means (shown here in grey for comparison).
That range was asymmetric and wider for the higher scenarios because
it included the uncertainty in carbon cycle climate feedbacks. The SRES
scenarios are based on the assumption of prescribed emissions, which
then translates to uncertainties in concentrations that propagate
through to uncertainties in the temperature response. The RCP sce-
narios assume prescribed concentrations. For scenarios that stabilize
(RCP2.6) that approach of constant fractional uncertainty underes-
timates the uncertainty and is no longer applicable, mainly because
internal variability has a larger relative contribution to the total uncer-
tainty (Good et al., 2013; Knutti and Sedláček, 2013). For the RCPs,
the carbon cycle climate feedback uncertainty is not included because
the simulations are driven by concentrations. Furthermore, there is no
clear evidence that distribution of CMIP5 global temperature changes
deviates from a normal distribution. For most other variables the shape
of the distribution is unclear, and standard deviations are simply used
as an indication of model spread, not representing a formal uncertainty
assessment.
Simulations with prescribed CO
2
emissions rather than concentrations
are only available for RCP8.5 (Figure 12.8b) and from MAGICC. The
projected temperature change in 2100 is slightly higher and the uncer-
tainty range is wider as a result of uncertainties in the carbon cycle
climate feedbacks. The CMIP5 range is consistent with the uncertainty
range given in AR4 for SRES A2 in 2100. Further details about emission
versus concentration driven simulations are given in Section 12.4.8.
In summary, the projected changes in global temperature for 2100 in
the RCP scenarios are very consistent with those obtained by CMIP3
for SRES in IPCC AR4 (see Section 12.4.9) when taking into account the
differences in scenarios. The likely uncertainty ranges provided here are
similar for RCP4.5 and RCP6.0 but narrower for RCP8.5 compared to
AR4. There was no scenario as low as RCP2.6 in AR4. The uncertainties
in global temperature projections have not decreased significantly in
CMIP5 (Knutti and Sedláček, 2013), but the assessed ranges cannot be
compared between AR4 and AR5. The main reason is that uncertain-
ties in carbon cycle feedbacks are not considered in the concentration
driven RCPs. In contrast, the likely range in AR4 included those. The
assessed likely ranges are therefore narrower for the high RCPs. The
differences in the projected warming are largely attributable to the dif-
ference in scenarios (Knutti and Sedláček, 2013), and the change in the
future and reference period, rather than to developments in modelling
since AR4. A detailed comparison between the SRES and RCP scenarios
and the CMIP3 and CMIP5 models is given in Section 12.4.9.
12.4.2 Pattern Scaling
12.4.2.1 Definition and Use
In this chapter we show geographical patterns of projected changes
in climate variables according to specific scenarios and time horizons.
Alternative scenarios and projection times can be inferred from those
shown by using some established approximation methods. This is espe-
cially the case for large-scale regional patterns of average temperature
and—with additional caveats—precipitation changes. In fact, ‘pattern
scaling’ is an approximation that has been explicitly suggested in the
description of the RCPs (Moss et al., 2010) as a method for deriving
impact-relevant regional projections for scenarios that have not been
simulated by global and regional climate models. It was first proposed
RCP2.6 RCP4.5 RCP6.0 RCP8.5
0
1
2
3
4
5
6
7
Concentration-driven
RCP2.6 RCP4.5 RCP6.0 RCP8.5
0
1
2
3
4
5
6
7
Emission-driven
CMIP5 models
90% range
66% range
median
Rogelj et al. 2012 Good et al. 2011
multimodel mean
-40 to +60% range
around mean
likely range
50th percentile
5-95th perc. range
Temperature increase in 2081-2100
relative to 1986 to 2005 (°C)
Temperature increase in 2081-2100
relative to 1986 to 2005 (°C)
a
b
Figure 12.8 | Uncertainty estimates for global mean temperature change in 2081–
2100 with respect to 1986–2005. Red crosses mark projections from individual CMIP5
models. Red bars indicate mean and 5 to 95% ranges based on CMIP5 (1.64 standard
deviations), which are considered as a likely range. Blue bars indicate 5 to 95% ranges
from the pulse response emulation of 21 models (Good et al., 2011a). Grey bars mark
the range from the mean of CMIP5 minus 40% to the mean +60%, assessed as likely in
AR4 for the SRES scenarios. The yellow bars show the median, 17 to 83% range and 5
to 95% range based on Rogelj et al. (2012). See also Figures 12.39 and 12.40.
1059
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Figure 12.9 | Surface air temperature change in 2081–2100 displayed as anomalies with respect to 1986–2005 for RCP4.5 from one ensemble member of each of the concen-
tration-driven models available in the CMIP5 archive.
Annual mean surface air temperature change (RCP4.5: 2081-2100)
by Santer et al. (1990) and revisited later by numerous studies (e.g.,
Huntingford and Cox, 2000). It relies on the existence of robust geo-
graphical patterns of change, emerging at the time when the response
to external forcings emerges from the noise, and persisting across the
length of the simulation, across different scenarios, and even across
models, modulated by the corresponding changes in global average
temperature. The robustness of temperature change patterns has
been amply documented from the original paper onward. An example
is given in Figure 12.9 for surface air temperature from each of the
CMIP5 models highlighting both similarities and differences between
the responses of different models. The precipitation pattern was shown
to scale linearly with global average temperature to a sufficient accu-
racy in CMIP3 models (Neelin et al., 2006) for this to be useful for
projections related to the hydrological cycle. Shiogama et al. (2010b)
find similar results with the caution that in the early stages of warming
aerosols modify the pattern. A more mixed evaluation can be found in
1060
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Good et al. (2012), where some land areas in the low latitudes exhibit a
nonlinear relation to global average temperature, but, largely, average
precipitation change over the remaining regions can be well approx-
imated by a grid-point specific linear function of global average tem-
perature change. It is in the latter quantity that the dependence of the
evolution of the change in time on the model (e.g., its climate sensitivi-
ty) and the forcing (e.g., the emission scenario) is encapsulated.
In analytical terms, it is assumed that the following relation holds:
C (t,x) = T
G
(t) c(x) + R (t, x)
where the symbol x identifies the geographic location (model grid
point or other spatial coordinates) and possibly the time of year (e.g.,
a June–July–August average). The index t runs along the length of the
forcing scenario of interest. T
G
(t) indicates global average temperature
change at time t under this scenario; c(x) is the time-invariant geo-
graphic pattern of change per 1°C global surface temperature change
for the variable of interest (which represents the forced component of
the change) and C (t,x) is the actual field of change for that variable
at the specific time t under this scenario. The R (t, x) is a residual term
and highlights the fact that pattern scaling cannot reconstruct model
behaviour with complete accuracy due to both natural variability and
because of limitations of the methodology discussed below. This way,
regionally and temporally differentiated results under different scenar-
ios or climate sensitivities can be approximated by the product of a
spatial pattern, constant over time, scenario and model characteristics,
and a time evolving global mean change in temperature. Model and
scenario dependence are thus captured through the global mean tem-
perature response, and simple climate models calibrated against fully
coupled climate models can be used to simulate the latter, at a great
saving in computational cost. The spatial pattern can be estimated
through the available coupled model simulations under the assump-
tion that it does not depend on the specific scenario(s) used.
The choice of the pattern in the studies available in the literature can be
as simple as the ensemble average field of change (across models and/
or across scenarios, for the coupled experiments available), normalized
by the corresponding change in global average temperature, choosing
a segment of the simulations when the signal has emerged from the
noise of natural variability from a baseline of reference (e.g., the last
20 years of the 21st century compared to pre-industrial or current cli-
mate) and taking the difference of two multi-decadal means. Similar
properties and results have been obtained using more sophisticated
multivariate procedures that optimize the variance explained by the
pattern (Holden and Edwards, 2010). The validity of this approximation
is discussed by Mitchell et al. (1999) and Mitchell (2003). Huntingford
and Cox (2000) evaluate the quality of the approximation for numer-
ous variables, showing that the technique performs best for temper-
ature, downward longwave radiation, relative humidity, wind speeds
and surface pressure while showing relatively larger limitations for
rainfall rate anomalies. Joshi et al. (2013) have recently shown that the
accuracy of the approximation, especially across models, is improved
by adding a second term, linear in the land–sea surface warming ratio,
another quantity that can be easily estimated from existing coupled
climate model simulations. There exist of course differences between
the patterns generated by different GCMs (documented for example
for CMIP3 in Watterson and Whetton, 2011b), but uncertainty can be
characterized, for example, by the inter-model spread in the pattern
c(x). Recent applications of the methodology to probabilistic future
projections have in fact sought to fully quantify errors introduced by
the approximation, on the basis of the available coupled model runs
(Harris et al., 2006).
Pattern scaling and its applications have been documented in IPCC
WGI Reports before (IPCC, 2001, Section 13.5.2.1; Meehl et al., 2007b,
Section 10.3.2). It has been used extensively for regional tempera-
ture and precipitation change projections, for example, Murphy et al.
(2007), (Watterson, 2008), Giorgi (2008), Harris et al. (2006, 2010), May
(2008a), Ruosteenoja et al. (2007), Räisänen and Ruokolainen (2006),
Cabre et al. (2010) and impact studies, for example, as described in
Dessai et al. (2005) and Fowler et al. (2007b). Recent studies have
focussed on patterns linked to warming at certain global average tem-
perature change thresholds (e.g., May, 2008a; Sanderson et al., 2011)
and patterns derived under the RCPs (Ishizaki et al., 2012).
There are basic limitations to this approach, besides a degradation of
its performance as the regional scale of interest becomes finer and in
the presence of regionally specific forcings. Recent work with MIROC3.2
(Shiogama et al., 2010a; Shiogama et al., 2010b) has revealed a depend-
ence of the precipitation sensitivity (global average precipitation change
per 1°C of global warming—see Figure 12.6) on the scenario, due to the
precipitation being more sensitive to carbon aerosols than WMGHGs.
In fact, there are significant differences in black and organic carbon
aerosol forcing between the emission scenarios investigated by Shiog-
ama et al. (2010a; 2010b). Levy II et al. (2013) confirm that patterns of
precipitation change are spatially correlated with the sources of aerosol
emissions, in simulations where the indirect effect is represented. This
is a behaviour that is linked to a more general limitation of pattern
scaling, which breaks down if aerosol forcing is significant. The effects
of aerosols have a regional nature and are thus dependent on the future
sources of pollution which are likely to vary geographically in the future
and are difficult to predict (May, 2008a). For example, Asian and North
American aerosol production are likely to have different time histories
and future projections. Schlesinger et al. (2000) extended the method-
ology of pattern scaling by isolating and recombining patterns derived
by dedicated experiments with a coupled climate model where sulphate
aerosols were increased for various regions in turn. More recently, in
an extension of pattern scaling into a probabilistic treatment of model,
scenario and initial condition uncertainties, Frieler et al. (2012) derived
joint probability distributions for regionally averaged temperature and
precipitation changes as linear functions of global average temperature
and additional predictors including regionally specific sulphate aerosol
and black carbon emissions.
Pattern scaling is less accurate for strongly mitigated stabilization
scenarios. This has been shown recently by May (2012), compar-
ing patterns of temperature change under a scenario limiting global
warming since pre-industrial times to 2°C and patterns produced by
a scenario that reaches 4.5°C of global average temperature change.
The limitations of pattern scaling in approximating changes while the
climate system approaches equilibrium have found their explanation in
Manabe and Wetherald (1980) and Mitchell et al. (1999). Both studies
point out that as the temperatures of the deep oceans reach equilibri-
1061
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
um (over multiple centuries) the geographical distribution of warming
changes as well, for example, showing a larger warming of the high
latitudes in the SH than in the earlier periods of the transient response,
relative to the global mean warming. More recently, Held et al. (2010)
showed how this slow warming pattern is in fact present during the
initial transient response of the system as well, albeit with much small-
er amplitude. Further, Gillett et al. (2011) show how in a simulation in
which emissions cease, regional temperatures and precipitation pat-
terns exhibit ongoing changes, even though global mean temperature
remains almost constant. Wu et al. (2010) showed that the global pre-
cipitation response shows a nonlinear response to strong mitigation
scenarios, with the hydrological cycle continuing to intensify even after
atmospheric CO
2
concentration, and thus global average temperature,
start decreasing. Regional nonlinear responses to mitigation scenari-
os of precipitation and sea surface temperatures (SSTs) are shown by
Chadwick et al. (2013).
Other areas where pattern scaling shows a lack of robustness are the
edges of polar ice caps and sea ice extent, where at an earlier time in
the simulation ice melts and regions of sharp gradient surface, while
later in the simulation, in the absence of ice, the gradient will become
less steep. Different sea ice representations in models also make the
location of such regions much less robust across the model ensembles
and the scenarios.
Pattern scaling has not been as thoroughly explored for quantities
other than average temperature and precipitation. Impact relevant
extremes, for example, seem to indicate a critical dependence on the
scale at which their changes are evaluated, with studies showing that
some aspects of their statistics change in a close-to-linear way with
mean temperature (Kharin et al., 2007; Lustenberger et al., 2013) while
others have documented the dependence of their changes on moments
of their statistical distribution other than the mean (Ballester et al.,
2010a), which would make pattern scaling inadequate.
12.4.2.2 Coupled Model Intercomparison Project Phase 5 Patterns
Scaled by Global Average Temperature Change
On the basis of CMIP5 simulations, we show geographical patterns
(Figure 12.10) of warming and precipitation change and indicate
measures of their variability across models and across RCPs. The pat-
terns are scaled to 1°C global mean surface temperature change above
the reference period 1986–2005 for 2081–2100 (first row) and for a
period of approximate stable temperature, 2181–2200 (thus excluding
RCP8.5, which does not stabilize by that time) (second row). Spatial
correlation of fields of temperature and precipitation change range
from 0.93 to 0.99 when considering ensemble means under different
RCPs. The lower values are found when computing correlation between
RCP2.6 and the higher RCPs, and may be related to the high mitigation
Precipitation scaled by global T (% per
o
C)Temperature scaled by global T (
o
C per
o
C)
Figure 12.10 | Temperature (left) and precipitation (right) change patterns derived from transient simulations from the CMIP5 ensembles, scaled to 1°C of global mean surface
temperature change. The patterns have been calculated by computing 20-year averages at the end of the 21st (top) and 22nd (bottom) centuries and over the period 1986–2005
for the available simulations under all RCPs, taking their difference (percentage difference in the case of precipitation) and normalizing it, grid-point by grid-point, by the cor-
responding value of global average temperature change for each model and scenario. The normalized patterns have then been averaged across models and scenarios. The colour
scale represents degrees Celsius (in the case of temperature) and percent (in the case of precipitation) per 1°C of global average temperature change. Stippling indicates where the
mean change averaged over all realizations is larger than the 95% percentile of the distribution of models. Zonal means of the geographical patterns are shown for each individual
model for RCP2.6 (blue), 4.5 (light blue), 6.0 (orange) and 8.5 (red). RCP8.5 is excluded from the stabilization figures. The RCP2.6 simulation of the FIO-ESM (First Institute of
Oceanography) model was excluded because it did not show any warming by the end of the 21st century, thus not complying with the method requirement that the pattern be
estimated at a time when the temperature change signal from CO
2
increase has emerged.
1062
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
enacted under RCP2.6 from early in the 21st century. Pattern corre-
lation varies between 0.91 and 0.98 for temperature and between
0.91 and 0.96 for precipitation when comparing patterns computed
by averaging and normalizing changes at the end of the 21st, 22nd
and 23rd centuries, with the largest value representing the correlation
between the patterns at the end of the 22nd and 23rd centuries, the
lowest representing the correlation between the pattern at the end
of the 21st and the pattern at the end of the 23rd century. The zonal
means shown to the side of each plot represent each model by one line,
colour coding the four different scenarios. They show good agreement
of models and scenarios over low and mid-latitudes for temperature,
but higher spread across models and especially across scenarios for the
areas subject to polar amplification, for which the previous discussion
about the sensitivity of the patterns to the sea ice edge may be rele-
vant. A comparison of the mean of the lines to their spread indicates
overall the presence of a strong mean signal with respect to the spread
of the ensemble. Precipitation shows an opposite pattern of inter-mod-
el spread, with larger variations in the low latitudes and around the
equator, and smaller around the high latitudes. Precipitation has also
a lower signal-to-noise ratio (measured as above by comparing the
ensemble mean change magnitude to the spread across models and
scenarios of these zonal mean averages).
As already mentioned, although we do not explicitly use pattern scaling
in the sections that follow, we consider it a useful approximation when
the need emerges to interpolate or extrapolate results to different sce-
narios or time periods, noting the possibility that the scaling may break
down at higher levels of global warming, and that the validity of the
approximation is limited to broad patterns of change, as opposed to
local scales. An important caveat is that pattern scaling only applies
to the climate response that is externally forced. The actual response
is a combination of forced change and natural variability, which is not
and should not be scaled up or down by the application of this tech-
nique, which becomes important on small spatial scales and shorter
time scales, and whose relative magnitude compared to the forced
component also depends on the variable (Hawkins and Sutton, 2009,
2011; Mahlstein et al., 2011; Deser et al., 2012a, 2012b; Mahlstein et
al., 2012) (see Section 11.2). One approach to produce projections that
include both components is to estimate natural variability separately,
scale the forced response and add the two.
12.4.3 Changes in Temperature and Energy Budget
12.4.3.1 Patterns of Surface Warming: Land–Sea Contrast,
Polar Amplification and Sea Surface Temperatures
Patterns of surface air temperature change for various RCPs show
widespread warming during the 21st century (Figure 12.11; see
Annex I for seasonal patterns). A key feature that has been present
throughout the history of coupled modelling is the larger warming over
land compared to oceans, which occurs in both transient and equilib-
rium climate change (e.g., Manabe et al., 1990). The degree to which
warming is larger over land than ocean is remarkably constant over
time under transient warming due to WMGHGs (Lambert and Chiang,
2007; Boer, 2011; Lambert et al., 2011) suggesting that heat capac-
ity differences between land and ocean do not play a major role in
the land–sea warming contrast (Sutton et al., 2007; Joshi et al., 2008,
2013). The phenomenon is predominantly a feature of the surface and
lower atmosphere (Joshi et al., 2008). Studies have found it occurs due
to contrasts in surface sensible and latent fluxes over land (Sutton et
al., 2007), land–ocean contrasts in boundary layer lapse rate changes
(Joshi et al., 2008), boundary layer relative humidity and associated
low-level cloud cover changes over land (Doutriaux-Boucher et al.,
2009; Fasullo, 2010) and soil moisture reductions (Dong et al., 2009;
Clark et al., 2010) under climate change. The land–sea warming con-
trast is also sensitive to aerosol forcing (Allen and Sherwood, 2010;
Joshi et al., 2013). Globally averaged warming over land and ocean
is identified separately in Table 12.2 for the CMIP5 models and the
ratio of land to ocean warming is likely in the range of 1.4 to 1.7,
consistent with previous studies (Lambert et al., 2011). The CMIP5 mul-
ti-model mean ratio is approximately constant from 2020 through to
2100 (based on an update of Joshi et al., 2008 from available CMIP5
models).
Amplified surface warming in Arctic latitudes is also a consistent fea-
ture in climate model integrations (e.g., Manabe and Stouffer, 1980).
This is often referred to as polar amplification, although numerous
studies have shown that under transient forcing, this is primarily an
Arctic phenomenon (Manabe et al., 1991; Meehl et al., 2007b). The
lack of an amplified transient warming response in high Southern polar
latitudes has been associated with deep ocean mixing, strong ocean
heat uptake and the persistence of the vast Antarctic ice sheet. In equi-
librium simulations, amplified warming occurs in both polar regions.
On an annual average, and depending on the forcing scenario (see
Table 12.2), the CMIP5 models show a mean Arctic (67.5°N to 90°N)
warming between 2.2 and 2.4 times the global average warming for
2081–2100 compared to 1986–2005. Similar polar amplification fac-
tors occurred in earlier coupled model simulations (e.g., Holland and
Bitz, 2003; Winton, 2006a). This factor in models is slightly higher
than the observed central value, but it is within the uncertainty of
the best estimate from observations of the recent past (Bekryaev et
al., 2010). The uncertainty is large in the observed factor because sta-
tion records are short and sparse (Serreze and Francis, 2006) and the
forced signal is contaminated by the noise of internal variability. By
contrast, model trends in surface air temperature are 2.5 to 5 times
higher than observed over Antarctica, but here also the observational
estimates have a very large uncertainty, so, for example, the CMIP3
ensemble mean is consistent with observations within error estimates
(Monaghan et al., 2008). Moreover, recent work suggests more wide-
spread current West Antarctic surface warming than previously esti-
mated (Bromwich et al., 2013).
The amplified Arctic warming in models has a distinct seasonal charac-
ter (Manabe and Stouffer, 1980; Rind, 1987; Holland and Bitz, 2003; Lu
and Cai, 2009; Kumar et al., 2010). Arctic amplification (defined as the
67.5 N° to 90°N warming compared to the global average warming
for 2081–2100 versus 1986–2005) peaks in early winter (November
to December) with a CMIP5 RCP4.5 multi-model mean warming for
67.5°N to 90°N exceeding the global average by a factor of more than
4. The warming is smallest in summer when excess heat at the Arctic
surface goes into melting ice or is absorbed by the ocean, which has
a relatively large thermal inertia. Simulated Arctic warming also has
a consistent vertical structure that is largest in the lower troposphere
1063
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
(e.g., Manabe et al., 1991; Kay et al., 2012). This is in agreement with
recent observations (Serreze et al., 2009; Screen and Simmonds, 2010)
but contrary to an earlier study that suggested a larger warming aloft
(Graversen et al., 2008). The discrepancy in observed vertical structure
may reflect inadequacies in data sets (Bitz and Fu, 2008; Grant et al.,
2008; Thorne, 2008) and sensitivity to the time period used for averag-
ing (see also Box 2.3).
As also discussed in Box 5.1, there are many mechanisms that con-
tribute to Arctic amplification, some of which were identified in early
modelling studies (Manabe and Stouffer, 1980). Feedbacks associat-
ed with changes in sea ice and snow amplify surface warming near
the poles (Hall, 2004; Soden et al., 2008; Graversen and Wang, 2009;
Kumar et al., 2010). The longwave radiation changes in the top of the
atmosphere associated with surface warming opposes surface warm-
ing at all latitudes, but less so in the Arctic (Winton, 2006a; Soden et
al., 2008). Rising temperature globally is expected to increase the hori-
Annual mean surface air temperature change
Figure 12.11 | Multi-model ensemble average of surface air temperature change (compared to 1986–2005 base period) for 2046–2065, 2081–2100, 2181–2200 for RCP2.6,
4.5, 6.0 and 8.5. Hatching indicates regions where the multi-model mean change is less than one standard deviation of internal variability. Stippling indicates regions where the
multi-model mean change is greater than two standard deviations of internal variability and where at least 90% of the models agree on the sign of change (see Box 12.1). The
number of CMIP5 models used is indicated in the upper right corner of each panel.
zontal latent heat transport by the atmosphere into the Arctic (Flan-
nery, 1984; Alexeev et al., 2005; Cai, 2005; Langen and Alexeev, 2007;
Kug et al., 2010), which warms primarily the lower troposphere. On
average, CMIP3 models simulate enhanced latent heat transport (Held
and Soden, 2006), but north of about 65°N, the sensible heat transport
declines enough to more than offset the latent heat transport increase
(Hwang et al., 2011). Increased atmospheric heat transport into the
Arctic and subsidence warming has been associated with a teleconnec-
tion driven by enhanced convection in the tropical western Pacific (Lee
et al., 2011). Ocean heat transport plays a role in the simulated Arctic
amplification, with both large late 20th century transport (Mahlstein
and Knutti, 2011) and increases over the 21st century (Hwang et al.,
2011; Bitz et al., 2012) associated with higher amplification. As noted
by Held and Soden (2006), Kay et al. (2012), and Alexeev and Jackson
(2012), diagnosing the role of various factors in amplified warming is
complicated by coupling in the system in which local feedbacks inter-
act with poleward heat transports.
1064
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Although models consistently exhibit Arctic amplification as global
mean temperatures rise, the multitude of physical processes described
above mean that they differ considerably in the magnitude. Previous
work has implicated variations across climate models in numerous fac-
tors including inversion strength (Boé et al., 2009a), ocean heat trans-
port (Holland and Bitz, 2003; Mahlstein and Knutti, 2011), albedo feed-
back (Winton, 2006a), longwave radiative feedbacks (Winton, 2006a)
and shortwave cloud feedback (Crook et al., 2011; Kay et al., 2012)
as playing a role in the across-model scatter in polar amplification.
The magnitude of amplification is generally higher in models with less
extensive late 20th century sea ice in June, suggesting that the initial
ice state influences the 21st century Arctic amplification. The pattern
of simulated Arctic warming is also associated with the initial ice state,
and in particular with the location of the winter sea ice edge (Holland
and Bitz, 2003; Räisänen, 2007; Bracegirdle and Stephenson, 2012).
This relationship has been suggested as a constraint on projected
Arctic warming (Abe et al., 2011; Bracegirdle and Stephenson, 2012),
although, in general, the ability of models to reproduce observed cli-
mate and its trends is not a sufficient condition for attributing high
confidence to the projection of future trends (see Section 9.8).
Minima in surface warming occur in the North Atlantic and Southern
Oceans under transient forcing in part due to deep ocean mixed layers
in those regions (Manabe et al., 1990; Xie et al., 2010). Trenberth and
Fasullo (2010) find that the large biases in the Southern Ocean energy
budget in CMIP3 coupled models negatively correlate with equilibrium
climate sensitivity (see Section 12.5.3), suggesting that an improved
mean state in the Southern Ocean is needed before warming there
can be understood. In the equatorial Pacific, warming is enhanced
in a narrow band which previous assessments have described as ‘El
Niño-like’, as may be expected from the projected decrease in atmos-
pheric tropical circulations (see Section 12.4.4). However, DiNezio et al.
(2009) highlight that the tropical Pacific warming in the CMIP3 models
is not ‘El Niño-like’ as the pattern of warming and associated tele-
connections (Xie et al., 2010; Section 12.4.5.2) is quite distinct from
that of an El Niño event. Instead the pattern is of enhanced equatorial
warming and is due to a meridional minimum in evaporative damping
on the equator (Liu et al., 2005) and ocean dynamical changes that can
be decoupled from atmospheric changes (DiNezio et al., 2009) (see
also further discussion in Section 12.4.7).
In summary, there is robust evidence over multiple generations of
models and high confidence in these large-scale warming patterns. In
the absence of a strong reduction in the Atlantic Meridional Overturn-
ing Circulation (AMOC), there is very high confidence that the Arctic
region is projected to warm most.
12.4.3.2 Zonal Average Atmospheric Temperature
Zonal temperature changes at the end of the 21st century show warm-
ing throughout the troposphere and, depending on the scenario, a mix
of warming and cooling in the stratosphere (Figure 12.12). The max-
imum warming in the tropical upper troposphere is consistent with
theoretical explanations and associated with a decline in the moist
adiabatic lapse rate of temperature in the tropics as the climate warms
(Bony et al., 2006). The northern polar regions also experience large
warming in the lower atmosphere, consistent with the mechanisms
discussed in Section 12.4.3.1. The tropospheric patterns are similar
to those in the TAR and AR4 with the RCP8.5 changes being up to
several degrees warmer in the tropics compared to the A1B changes
appearing in the AR4. Similar tropospheric patterns appear in the RCP
2.6 and 4.5 changes, but with reduced magnitudes, suggesting some
degree of scaling with forcing change in the troposphere, similar to
behaviour discussed in the AR4 and Section 12.4.2. The consistency of
tropospheric patterns over multiple generations of models indicates
high confidence in these projected changes.
In the stratosphere, the models show similar tropical patterns of
change, with magnitudes differing according to the degree of cli-
mate forcing. Substantial differences appear in polar regions. In the
north, RCP8.5 and 4.5 yield cooling, though it is more significant in
the RCP8.5 ensemble. In contrast, RCP2.6 shows warming, albeit weak
and with little significance. In the southern polar region, RCP 2.6 and
4.5 both show significant warming, and RCP8.5 is the outlier, with sig-
nificant cooling. The polar stratospheric warming, especially in the SH,
is similar to that found by Butchart et al. (2010) and Meehl et al. (2012)
in GCM simulations that showed effects of ozone recovery in deter-
mining the patterns (Baldwin et al., 2007; Son et al., 2010). Eyring et
al. (2013) find behaviour in the CMIP5 ensemble both for models with
and without interactive chemistry that supports the contention that
the polar stratospheric changes in Figure 12.12 are strongly influenced
by ozone recovery. Overall, the stratospheric temperature changes do
not exhibit pattern scaling with global temperature change and are
dependent on ozone recovery.
Away from the polar stratosphere, there is physical and pattern consist-
ency in temperature changes between different generations of models
assessed here and in the TAR and AR4. The consistency is especially clear
in the northern high latitudes and, coupled with physical understanding,
indicates that some of the greatest warming is very likely to occur here.
There is also consistency across generations of models in relatively large
warming in the tropical upper troposphere. Allen and Sherwood (2008)
and Johnson and Xie (2010) have presented dynamic and thermody-
namic arguments, respectively, for the physical robustness of the tropi-
cal behaviour. However, there remains uncertainty about the magnitude
of warming simulated in the tropical upper troposphere because large
observational uncertainties and contradictory analyses limit a confident
assessment of model accuracy in simulating temperature trends in the
tropical upper troposphere (Section 9.4.1.4.2). The combined evidence
indicates that relatively large warming in the tropical upper troposphere
is likely, but with medium confidence.
12.4.3.3 Temperature Extremes
As the climate continues to warm, changes in several types of tem-
perature extremes have been observed (Donat et al., 2013), and are
expected to continue in the future in concert with global warming
(Seneviratne et al., 2012). Extremes occur on multiple time scales, from
a single day or a few consecutive days (a heat wave) to monthly and
seasonal events. Extreme temperature events are often defined by
indices (see Box 2.4 for the common definitions used), for example,
percentage of days in a year when maximum temperature is above the
90th percentile of a present day distribution or by long period return
values. Although changes in temperature extremes are a very robust
1065
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
signature of anthropogenic climate change (Seneviratne et al., 2012),
the magnitude of change and consensus among models varies with
the characteristics of the event being considered (e.g., time scale, mag-
nitude, duration and spatial extent) as well as the definition used to
describe the extreme.
Since the AR4 many advances have been made in establishing global
observed records of extremes (Alexander et al., 2006; Perkins et al.,
2012; Donat et al., 2013) against which models can be evaluated to
give context to future projections (Sillmann and Roeckner, 2008; Alex-
ander and Arblaster, 2009). Numerous regional assessments of future
changes in extremes have also been performed and a comprehensive
summary of these is given in Seneviratne et al. (2012). Here we sum-
marize the key findings from this report and assess updates since then.
It is virtually certain that there will be more hot and fewer cold extremes
as global temperature increases (Caesar and Lowe, 2012; Orlowsky
and Seneviratne, 2012; Sillmann et al., 2013), consistent with previous
assessments (Solomon et al., 2007; Seneviratne et al., 2012). Figure
12.13 shows multi-model mean changes in the absolute temperature
indices of the coldest day of the year and the hottest day of the year
and the threshold-based indices of frost days and tropical nights from
the CMIP5 ensemble (Sillmann et al., 2013). A robust increase in warm
temperature extremes and decrease in cold temperature extremes
is found at the end of the 21st century, with the magnitude of the
changes increasing with increased anthropogenic forcing. The coldest
night of the year undergoes larger increases than the hottest day in
the globally averaged time series (Figure 12.13b and d). This tenden-
cy is consistent with the CMIP3 model results shown in Figure 12.13,
which use different models and the SRES scenarios (see Seneviratne
et al. (2012) for earlier CMIP3 results). Similarly, increases in the fre-
quency of warm nights are greater than increases in the frequency
of warm days (Sillmann et al., 2013). Regionally, the largest increases
in the coldest night of the year are projected in the high latitudes of
Figure 12.12 | CMIP5 multi-model changes in annual mean zonal mean temperature in the atmosphere and ocean relative to 1986–2005 for 2081–2100 under the RCP2.6 (left),
RCP4.5 (centre) and RCP8.5 (right) forcing scenarios. Hatching indicates regions where the multi-model mean change is less than one standard deviation of internal variability.
Stippling indicates regions where the multi-model change mean is greater than two standard deviations of internal variability and where at least 90% of the models agree on the
sign of change (see Box 12.1).
1066
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
the NH under the RCP8.5 scenario (Figure 12.13a). The subtropics and
mid-latitudes exhibit the greatest projected changes in the hottest day
of the year, whereas changes in tropical nights and the frequency of
warm days and warm nights are largest in the tropics (Sillmann et al.,
2013). The number of frost days declines in all regions while significant
increases in tropical nights are seen in southeastern North America, the
Mediterranean and central Asia.
It is very likely that, on average, there will be more record high than
record cold temperatures in a warmer average climate. For example,
Meehl et al. (2009) find that the current ratio of 2 to 1 for record daily
high maxima to low minima over the USA becomes approximately 20
to 1 by the mid-21st century and 50 to 1 by late century in their model
simulation of the SRES A1B scenario. However, even at the end of the
century daily record low minima continue to be broken, if in a small
number, consistent with Kodra et al. (2011), who conclude that cold
extremes will continue to occur in a warmer climate, even though their
frequency will decline.
It is also very likely that heat waves, defined as spells of days with
temperature above a threshold determined from historical climatology,
will occur with a higher frequency and duration, mainly as a direct
consequence of the increase in seasonal mean temperatures (Barnett
et al., 2006; Ballester et al., 2010a, 2010b; Fischer and Schär, 2010).
Changes in the absolute value of temperature extremes are also very
likely and expected to regionally exceed global temperature increases
by far, with substantial changes in hot extremes projected even for
moderate (<2.5°C above present day) average warming levels (Clark
et al., 2010; Diffenbaugh and Ashfaq, 2010). These changes often differ
from the mean temperature increase, as a result of changes in variabili-
ty and shape of the temperature distribution (Hegerl et al., 2004; Meehl
and Tebaldi, 2004; Clark et al., 2006). For example, summer tempera-
ture extremes over central and southern Europe are projected to warm
substantially more than the corresponding mean local temperatures as
a result of enhanced temperature variability at interannual to intrasea-
sonal time scales (Schär et al., 2004; Clark et al., 2006; Kjellstrom et
al., 2007; Vidale et al., 2007; Fischer and Schär, 2009, 2010; Nikulin et
al., 2011; Fischer et al., 2012a). Several recent studies have also argued
that the probability of occurrence of a Russian heat wave at least as
severe as the one in 2010 increases substantially (by a factor of 5 to
10 by the mid-century) along with increasing mean temperatures and
enhanced temperature variability (Barriopedro et al., 2011; Dole et al.,
2011).
Since the AR4, an increased understanding of mechanisms and feed-
backs leading to projected changes in extremes has been gained
(Seneviratne et al., 2012). Climate models suggest that hot extremes
are amplified by soil moisture-temperature feedbacks (Seneviratne et
al., 2006; Diffenbaugh et al., 2007; Lenderink et al., 2007; Vidale et
al., 2007; Fischer and Schär, 2009; Fischer et al., 2012a) in northern
mid-latitude regions as the climate warms, consistent with previous
assessments. Changes in temperature extremes may also be impacted
by changes in land–sea contrast, with Watterson et al. (2008) show-
ing an amplification of southern Australian summer warm extremes
over the mean due to anomalous temperature advection from warmer
continental interiors. The largest increases in the magnitude of warm
extremes are simulated over mid-latitude continental areas, consistent
with the drier conditions, and the associated reduction in evaporative
cooling from the land surface projected over these areas (Kharin et al.,
2007). The representation of the latter constitutes a major source of
model uncertainty for projections of the absolute magnitude of tem-
perature extremes (Clark et al., 2010; Fischer et al., 2011).
Winter cold extremes also warm more than the local mean temper-
ature over northern high latitudes (Orlowsky and Seneviratne, 2012;
Sillmann et al., 2013) as a result of reduced temperature variability
related to declining snow cover (Gregory and Mitchell, 1995; Kjellstrom
et al., 2007; Fischer et al., 2011) and decreases in landsea contrast
(de Vries et al., 2012). Changes in atmospheric circulation, induced by
remote surface heating can also modify the temperature distribution
(Haarsma et al., 2009). Sillmann and Croci-Maspoli (2009) note that
cold winter extremes over Europe are in part driven by atmospheric
blocking and changes to these blocking patterns in the future lead to
changes in the frequency and spatial distribution of cold temperature
extremes as global temperatures increase. Occasional cold winters will
continue to occur (Räisänen and Ylhaisi, 2011).
Human discomfort, morbidity and mortality during heat waves depend
not only on temperature but also specific humidity. Heat stress, defined
as the combined effect of temperature and humidity, is expected to
increase along with warming temperatures and dominates the local
decrease in summer relative humidity due to soil drying (Diffenbaugh
et al., 2007; Fischer et al., 2012b; Dunne et al., 2013). Areas with abun-
dant atmospheric moisture availability and high present-day temper-
atures such as Mediterranean coastal regions are expected to experi-
ence the greatest heat stress changes because the heat stress response
scales with humidity which thus becomes increasingly important to
heat stress at higher temperatures (Fischer and Schär, 2010; Sherwood
and Huber, 2010; Willett and Sherwood, 2012). For some regions, sim-
ulated heat stress indicators are remarkably robust, because those
models with stronger warming simulate a stronger decrease in atmos-
pheric relative humidity (Fischer and Knutti, 2013).
Changes in rare temperature extremes can be assessed using extreme
value theory based techniques (Seneviratne et al., 2012). Kharin et
al. (2007), in an analysis of CMIP3 models, found large increases
in the 20-year return values of the annual maximum and minimum
daily averaged surface air temperatures (i.e., the size of an event
that would be expected on average once every 20 years, or with a
5% chance every year) with larger changes over land than ocean.
Figure 12.14 displays the end of 21st century change in the magni-
tude of these rare events from the CMIP5 models in the RCP2.6, 4.5
and 8.5 scenarios (Kharin et al., 2013). Comparison to the changes in
summer mean temperature shown in Figure AI.5 and A1.7 of Annex
I Supplementary Material reveals that rare high temperature events
are projected to change at rates similar to or slightly larger than the
summertime mean temperature in many land areas. However, in much
of Northern Europe 20-year return values of daily high temperatures
are projected to increase 2°C or more than JJA mean temperatures
under RCP8.5, consistent with previous studies (Sterl et al., 2008;
Orlowsky and Seneviratne, 2012). Rare low temperature events are
projected to experience significantly larger increases than the mean
in most land regions, with a pronounced effect at high latitudes. Twen-
ty-year return values of cold extremes increase significantly more than
1067
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Frost Days (FD)
f)
−25
−20
−15
−10
−5
0
5
1960 1980 2000 2020 2040 2060 2080 2100
Year
−25
−20
−15
−10
−5
0
5
(days)
Warmest daily Tmax (TXx)
d)
1960 1980 2000 2020 2040 2060 2080 2100
Year
−2
0
2
4
6
8
−2
0
2
4
6
8
(
o
C)
−2
0
2
4
6
8
historical
RCP2.6
RCP4.5
RCP8.5
CMIP3 B1 CMIP3 A1B CMIP3 A2
historical
RCP2.6
RCP4.5
RCP8.5
CMIP3 B1 CMIP3 A1B CMIP3 A2
historical
RCP2.6
RCP4.5
RCP8.5
CMIP3 B1 CMIP3 A1B CMIP3 A2
1960 1980 2000 2020 2040 2060 2080 2100
Year
−2
0
2
4
6
8
Coldest daily Tmin (TNn)
b)
(
o
C)
Tropical Nights (TR)
h)
0
20
40
60
1960 1980 2000 2020 2040 2060 2080 2100
Year
0
20
40
60
(days)
historical
RCP2.6
RCP4.5
RCP8.5
CMIP3 B1 CMIP3 A1B CMIP3 A2
18
18
18
18
Figure 12.13 | CMIP5 multi-model mean geographical changes (relative to a 1981–2000 reference period in common with CMIP3) under RCP8.5 and 20-year smoothed time
series for RCP2.6, RCP4.5 and RCP8.5 in the (a, b) annual minimum of daily minimum temperature, (c, d) annual maximum of daily maximum temperature, (e, f) frost days (number
of days below 0°C) and (g, h) tropical nights (number of days above 20°C). White areas over land indicate regions where the index is not valid. Shading in the time series represents
the interquartile ensemble spread (25th and 75th quantiles). The box-and-whisker plots show the interquartile ensemble spread (box) and outliers (whiskers) for 11 CMIP3 model
simulations of the SRES scenarios A2 (orange), A1B (cyan), and B1 (purple) globally averaged over the respective future time periods (2046–2065 and 2081–2100) as anomalies
from the 1981–2000 reference period. Stippling indicates grid points with changes that are significant at the 5% level using a Wilcoxon signed-ranked test. (Updated from Sillmann
et al. (2013), excluding the FGOALS-s2 model.)
1068
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
winter mean temperature changes, particularly over parts of North
America and Europe. Kharin et al. (2013) concluded from the CMIP5
models that it is likely that in most land regions a current 20 year max-
imum temperature event is projected to become a one-in-two-year
event by the end of the 21st century under the RCP4.5 and RCP8.5
scenarios, except for some regions of the high latitudes of the NH
where it is likely to become a one-in-five-year event (see also Senevi-
ratne et al. (2012) Figure 3.5). Current 20-year minimum temperature
events are projected to become exceedingly rare, with return periods
likely increasing to more than 100 years in almost all locations under
RCP8.5 (Kharin et al., 2013). Section 10.6.1.1 notes that a number of
detection and attribution studies since SREX suggest that the model
changes may tend to be too large for warm extremes and too small
for cold extremes and thus these likelihood statements are somewhat
less strongly stated than a direct interpretation of model output and
its ranges. The CMIP5 analysis shown in Figure 12.14 reinforces this
assessment of large changes in the frequency of rare events, particu-
larly in the RCP8.5 scenario (Kharin et al., 2013).
There is high consensus among models in the sign of the future change
in temperature extremes, with recent studies confirming this conclu-
sion from the previous assessments (Tebaldi et al., 2006; Meehl et al.,
2007b; Orlowsky and Seneviratne, 2012; Seneviratne et al., 2012; Sill-
mann et al., 2013). However, the magnitude of the change remains
uncertain owing to scenario and model (both structural and parame-
ter) uncertainty (Clark et al., 2010) as well as internal variability. These
uncertainties are much larger than corresponding uncertainties in the
magnitude of mean temperature change (Barnett et al., 2006; Clark et
al., 2006; Fischer and Schär, 2010; Fischer et al., 2011).
Figure 12.14 | The CMIP5 multi-model median change in 20-year return values of annual warm temperature extremes (left-hand panels) and cold temperature extremes (right-
hand panels) as simulated by CMIP5 models in 2081–2100 relative to 1986–2005 in the RCP2.6 (top), RCP4.5 (middle panels), and RCP8.5 (bottom) experiments.
1069
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
12.4.3.4 Energy Budget
Anthropogenic or natural perturbations to the climate system produce
RFs that result in an imbalance in the global energy budget at the
top of the atmosphere (TOA) and affect the global mean temperature
(Section 12.3.3). The climate responds to a change in RF on multiple
time scales and at multiyear time scales the energy imbalance (i.e.,
the energy heating or cooling the Earth) is very close to the ocean
heat uptake due to the much lower thermal inertia of the atmosphere
and the continental surfaces (Levitus et al., 2005; Knutti et al., 2008a;
Murphy et al., 2009; Hansen et al., 2011). The radiative responses of
the fluxes at TOA are generally analysed using the forcing-feedback
framework and are presented in Section 9.7.2.
CMIP5 models simulate a small increase of the energy imbalance at
the TOA over the 20th century (see Box 3.1, Box 9.2 and Box 13.1). The
future evolution of the imbalance is very different depending on the
scenario (Figure 12.15a): for RCP8.5 it continues to increase rapidly,
much less for RCP6.0, it is almost constant for RCP4.5 and decreases
for RCP2.6. This latter negative trend reveals the quasi-stabilization
characteristic of RCP2.6. (In a transient scenario simulation, the TOA
imbalance is always less than the RF because of the slow rate of ocean
heat uptake.)
The rapid fluctuations that are simulated during the 20th century
originate from volcanic eruptions that are prescribed in the models
(see Section 12.3.2). These aerosols reflect solar radiation and thus
decrease the amount of SW radiation absorbed by the Earth (Figure
12.15c). The minimum of shortwave (SW) radiation absorbed by the
Earth during the period 1960–2000 is due mainly to two factors: a
sequence of volcanic eruptions and an increase of the reflecting aer-
osol burden due to human activities (see Sections 7.5, 8.5 and 9.4.6).
During the 21st century, the absorbed SW radiation monotonically
increases for the RCP8.5 scenario, and increases and subsequently
stabilizes for the other scenarios, consistent with what has been pre-
viously obtained with CMIP3 models and SRES scenarios (Trenberth
and Fasullo, 2009). The two main contributions to the SW changes are
the change of clouds (see Section 12.4.3.5) and the change of the cry-
osphere (see Section 12.4.6) at high latitudes. In the longwave (LW)
domain (Figure 12.15b), the net flux at TOA represents the opposite of
the flux that is emitted by the Earth’s surface and atmosphere toward
space, i.e., a negative anomaly represents an increase of the emitted
(W m
-2
)
Figure 12.15 | Time series of global and annual multi-model mean (a) net total radiation anomaly at the top of the atmosphere (TOA), (b) net longwave radiation anomaly at
the TOA and (c) net shortwave radiation anomaly at the TOA from the CMIP5 concentration-driven experiments for the historical period (black) and the four RCP scenarios. All the
fluxes are positive downward and units are W m
–2
. The anomalies are calculated relative to the 1900–1950 base period as this is a common period to all model experiments with
few volcanic eruptions and relatively small trends. One ensemble member is used for each individual CMIP5 model and the ± standard deviation across the distribution of individual
models is shaded.
Figure 12.16 | Multi-model CMIP5 average changes in annual mean (left) net total radiation anomaly at the top of the atmosphere (TOA), (middle) net longwave radiation
anomaly at the TOA and (right) net shortwave radiation anomaly at the TOA for the RCP4.5 scenario averaged over the periods 2081–2100. All fluxes are positive downward, units
are W m
–2
. The net radiation anomalies are computed with respect to the 1900–1950 base period. Hatching indicates regions where the multi-model mean change is less than one
standard deviation of internal variability. Stippling indicates regions where the multi-model mean change is greater than two standard deviations of internal variability and where
at least 90% of models agree on the sign of change (see Box 12.1).
1070
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
LW radiation. The LW net flux depends mainly on two factors: the sur-
face temperature and the magnitude of the greenhouse effect of the
atmosphere. During the 20th century, the rapid fluctuations of LW radi-
ation are driven by volcanic forcings, which decrease the absorbed SW
radiation, surface temperature, and the LW radiation emitted by the
Earth toward space. During the period 1960–2000, the fast increase of
GHG concentrations also decreases the radiation emitted by the Earth.
In response to this net heating of the Earth, temperatures warm and
thereby increase emitted LW radiation although the change of the tem-
perature vertical profile, water vapour, and cloud properties modulate
this response (e.g., Bony et al., 2006; Randall et al., 2007).
12.4.3.5 Clouds
This section provides a summary description of future changes in
clouds and their feedbacks on climate. A more general and more pre-
cise description and assessment of the role of clouds in the climate
system is provided in Chapter 7, in particular Section 7.2 for cloud pro-
cesses and feedbacks and Section 7.4 for aerosol–cloud interactions.
Cloud feedbacks and adjustments are presented in Section 7.2.5 and a
synthesis is provided in Section 7.2.6. Clouds are a major component
of the climate system and play an important role in climate sensitiv-
ity (Cess et al., 1990; Randall et al., 2007), the diurnal temperature
range (DTR) over land (Zhou et al., 2009), and land–sea contrast (see
Section 12.4.3.1). The observed global mean cloud RF is about –20 W
m
–2
(Loeb et al., 2009) (see Section 7.2.1), that is, clouds have a net
cooling effect. Current GCMs simulate clouds through various complex
parameterizations (see Section 7.2.3), and cloud feedback is a major
source of the spread of the climate sensitivity estimate (Soden and
Held, 2006; Randall et al., 2007; Dufresne and Bony, 2008) (see Section
9.7.2).
Under future projections the multi-model pattern of total cloud
amount shows consistent decreases in the subtropics, in conjunction
with a decrease of the relative humidity there, and increases at high
latitudes. Another robust pattern is an increase in cloud cover at all
latitudes in the vicinity of the tropopause, a signature of the increase of
the altitude of high level clouds in convective regions (Wetherald and
Manabe, 1988; Meehl et al., 2007b; Soden and Vecchi, 2011; Zelinka
et al., 2012). Low-level clouds were identified as a primary cause of
inter-model spread in cloud feedbacks in CMIP3 models (Bony and
Dufresne, 2005; Webb et al., 2006; Wyant et al., 2006). Since AR4, these
results have been confirmed along with the positive feedbacks due to
high level clouds in the CMIP3 or CFMIP models (Zelinka and Hart-
mann, 2010; Soden and Vecchi, 2011; Webb et al., 2013) and CMIP5
models (Vial et al., 2013). Since AR4, the response of clouds has been
partitioned in a direct or ‘rapid’ response of clouds to CO
2
and a ‘slow’
response of clouds to the surface temperature increase (i.e., the usual
feedback response) (Gregory and Webb, 2008). The radiative effect of
clouds depends mainly on their fraction, optical depth and temper-
ature. The contribution of these variables to the cloud feedback has
been quantified for the multi-model CMIP3 (Soden and Vecchi, 2011)
and CFMIP1 database (Zelinka et al., 2012). These findings are con-
sistent with the radiative changes obtained with the CMIP5 models
(Figure 12.16) and may be summarized as follows (see Section 7.2.5
for more details).
The dominant contributor to the SW cloud feedback is the change in
cloud fraction. The reduction of cloud fraction between 50°S and 50°N,
except along the equator and the eastern part of the ocean basins
(Figure 12.17), contributes to an increase in the absorbed solar radi-
ation (Figure 12.16c). Physical mechanisms and the role of different
parameterizations have been proposed to explain this reduction of
low-level clouds (Zhang and Bretherton, 2008; Caldwell and Breth-
erton, 2009; Brient and Bony, 2013; Webb et al., 2013). Poleward of
50°S, the cloud fraction and the cloud optical depth increases, thereby
increasing cloud reflectance. This leads to a decrease of solar absorp-
tion around Antarctica where the ocean is nearly ice free in summer
(Figure 12.16c). However, there is low confidence in this result because
GCMs do not reproduce the nearly 100% cloud cover observed there
and the negative feedback could be overestimated (Trenberth and
Fasullo, 2010) or, at the opposite, underestimated because the cloud
optical depth simulated by models is biased high there (Zelinka et al.,
2012).
In the LW domain, the tropical high cloud changes exert the dominant
effect. A lifting of the cloud top with warming is simulated consistently
across models (Meehl et al., 2007b) which leads to a positive feed-
back whereby the LW emissions from high clouds decrease as they
cool (Figure 12.16b). The dominant driver of this effect is the increase
of tropopause height and physical explanations have been proposed
(Hartmann and Larson, 2002; Lorenz and DeWeaver, 2007; Zelinka
Figure 12.17 | CMIP5 multi-model changes in annual mean total cloud fraction (in %) relative to 1986–2005 for 2081–2100 under the RCP2.6 (left), RCP4.5 (centre) and RCP8.5
(right) forcing scenarios. Hatching indicates regions where the multi-model mean change is less than one standard deviation of internal variability. Stippling indicates regions where
the multi-model mean change is greater than two standard deviations of internal variability and where 90% of the models agree on the sign of change (see Box 12.1). The number
of CMIP5 models used is indicated in the upper right corner of each panel.
1071
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
and Hartmann, 2010). Although the decrease in cloudiness generally
increases outgoing longwave radiation and partly offsets the effect of
cloud rising, the net effect is a consistent positive global mean LW
cloud feedback across CMIP and CFMIP models. Global mean SW cloud
feedbacks range from slightly negative to strongly positive (Soden and
Vecchi, 2011; Zelinka et al., 2012), with an inter-model spread in net
cloud feedback being mainly attributable to low-level cloud changes.
In summary, both the multi-model mean and the inter-model spread of
the cloud fraction and radiative flux changes simulated by the CMIP5
models are consistent with those previously obtained by the CMIP3
models. These include decreases in cloud amount in the subtropics,
increases at high latitudes and increases in the altitude of high level
clouds in convective regions. Many of these changes have been under-
stood primarily as responses to large-scale circulation changes (see
Section 7.2.6).
12.4.4 Changes in Atmospheric Circulation
Projected changes in energy and water cycles couple with changes in
atmospheric circulation and mass distribution. Understanding this cou-
pling is necessary to assess physical behaviour underlying projected
changes, particularly at regional scales, revealing why changes occur
and the realism of the changes. The focus in this section is on atmos-
pheric circulation behaviour that CMIP5 GCMs resolve well. Thus, the
section includes discussion of extratropical cyclones but not tropical
cyclones: extratropical cyclones are fairly well resolved by most CMIP5
GCMs, whereas tropical cyclones are not, requiring resolutions finer
than used by the large majority of CMIP5 GCMs (see Section 9.5.4.3).
Detailed discussion of tropical cyclones appears in Section 14.6.1
(see also Section 11.3.2.5.3 for near term changes and Section 3.4.4
in Seneviratne et al. (2012)). Regional detail concerning extratropical
storm tracks, including causal processes, appears in Section 14.6.2
(see also Section 11.3.2.4 for near-term changes and Seneviratne et al.
(2012) for an assessment of projected changes related to weather and
climate extremes).
12.4.4.1 Mean Sea Level Pressure and Upper-Air Winds
Sea level pressure gives an indication of surface changes in atmos-
pheric circulation (Figure 12.18). As in previous assessments, a robust
feature of the pattern of change is a decrease in high latitudes and
increases in the mid-latitudes, associated with poleward shifts in the
SH mid-latitude storm tracks (Section 12.4.4.3) and positive trends
in the annular modes (Section 14.5) as well as an expansion of the
Hadley Cell (Section 12.4.4.2). Similar patterns of sea level pressure
change are found in observed trends over recent decades, suggest-
ing an already detectable change (Gillett and Stott, 2009; Section
10.3.3.4), although the observed patterns are influenced by both natu-
ral and anthropogenic forcing as well as internal variability and the
relative importance of these influences is likely to change in the future.
Internal variability has been found to play a large role in uncertainties
of future sea level pressure projections, particularly at higher latitudes
(Deser et al., 2012a).
In boreal winter, decreases of sea level pressure over NH high lati-
tudes are slightly weaker in the CMIP5 ensemble compared to previous
assessments, consistent with Scaife et al. (2012) and Karpechko and
Manzini (2012), who suggest that improvements in the representation
of the stratosphere can influence this pattern. In austral summer, the
SH projections are impacted by the additional influence of stratospher-
ic ozone recovery (see Section 11.3.2.4.2) which opposes changes due
to GHGs. Under the weaker GHG emissions of RCP2.6, decreases in sea
level pressure over the SH mid-latitudes and increases over SH high
latitudes are consistent with expected changes from ozone recovery
(Arblaster et al., 2011; McLandress et al., 2011; Polvani et al., 2011). For
Figure 12.18 | CMIP5 multi-model ensemble average of December, January and February (DJF, top row) and June, July and August (JJA, bottom row) mean sea level pressure
change (2081–2100 minus 1986–2005) for, from left to right, RCP2.6, 4.5 and 8.5. Hatching indicates regions where the multi-model mean change is less than one standard
deviation of internal variability. Stippling indicates regions where the multi-model mean change is greater than two standard deviations of internal variability and where at least
90% of models agree on the sign of change (see Box 12.1).
1072
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
all other RCPs, the magnitude of SH extratropical changes scales with
the RF, as found in previous model ensembles (Paeth and Pollinger,
2010; Simpkins and Karpechko, 2012).
Large increases in seasonal sea level pressure are also found in regions
of sub-tropical drying such as the Mediterranean and northern Africa
in DJF and Australia in JJA. Projected changes in the tropics are less
consistent across the models; however, a decrease in the eastern equa-
torial Pacific and increase over the maritime continent, associated with
a weakening of the Walker Circulation (Vecchi and Soden, 2007; Power
and Kociuba, 2011b), is found in all RCPs.
Future changes in zonal and annual mean zonal winds (Figure 12.19)
are seen throughout the atmosphere with stronger changes in higher
RCPs. Large increases in winds are evident in the tropical stratosphere
and a poleward shift and intensification of the SH tropospheric jet is
seen under RCP4.5 and RCP8.5, associated with an increase in the
SH upper tropospheric meridional temperature gradient (Figure 12.12)
(Wilcox et al., 2012). In the NH, the response of the tropospheric jet
is weaker and complicated by the additional thermal forcing of polar
amplification (Woollings, 2008). Barnes and Polvani (2013) evaluate
changes in the annual mean mid-latitude jets in the CMIP5 ensemble,
finding consistent poleward shifts in both hemispheres under RCP8.5
for the end of the 21st century. In the NH, the poleward shift is ~1°,
similar to that found for the CMIP3 ensemble (Woollings and Black-
burn, 2012). In the SH, the annual mean mid-latitude jet shifts pole-
ward by ~2° under RCP8.5 at the end of the 21st century in the CMIP5
multi-model mean (Barnes and Polvani, 2013), with a similar shift of
1.5° in the surface westerlies (Swart and Fyfe, 2012). A strengthen-
ing of the SH surface westerlies is also found under all RCPs except
RCP2.6 (Swart and Fyfe, 2012), with largest changes in the Pacific
basin (Bracegirdle et al., 2013). In austral summer, ozone recovery off-
sets changes in GHGs to some extent, with a weak reversal of the jet
shift found in the multi-model mean under the low emissions scenario
of RCP2.6 (Swart and Fyfe, 2012) and weak or poleward shifts in other
RCPs (Swart and Fyfe, 2012; Wilcox et al., 2012). Eyring et al. (2013)
note the sensitivity of the CMIP5 SH summertime circulation changes
to both the strength of the ozone recovery (simulated by some models
interactively) and the rate of GHG increases.
Although the poleward shift of the tropospheric jets are robust across
models and likely under increased GHGs, the dynamical mechanisms
behind these projections are still not completely understood and have
been explored in both simple and complex models (Chen et al., 2008;
Lim and Simmonds, 2009; Butler et al., 2010). The shifts are associated
with a strengthening in the upper tropospheric meridional temperature
gradient (Wilcox et al., 2012) and hypotheses for associated changes
in planetary wave activity and/or synoptic eddy characteristics that
impact on the position of the jet have been put forward (Gerber et
al., 2012). Equatorward biases in the position of the SH jet (Section
9.5.3.2), while somewhat improved over similar biases in the CMIP3
models (Kidston and Gerber, 2010) still remain, limiting our confidence
in the magnitude of future changes.
In summary, poleward shifts in the mid-latitude jets of about 1 to 2
degrees latitude are likely at the end of the 21st century under RCP8.5
in both hemispheres (medium confidence) with weaker shifts in the NH
and under lower emission scenarios. Ozone recovery will likely weaken
the GHG-induced changes in the SH extratropical circulation in austral
summer.
12.4.4.2 Planetary-Scale Overturning Circulations
Large-scale atmospheric overturning circulations and their interaction
with other atmospheric mechanisms are significant in determining trop-
ical climate and regional changes in response to enhanced RF. Observed
Figure 12.19 | Coupled Model Intercomparison Project Phase 5 (CMIP5) multi-model ensemble average of zonal and annual mean wind change (2081–2100 minus 1986–2005)
for, from left to right, Representative Concentration Pathway 2.6 (RCP2.6), 4.5 and 8.5. Black contours represent the multi-model average for the 1986–2005 base period. Hatching
indicates regions where the multi-model mean change is less than one standard deviation of internal variability. Stippling indicates regions where the multi-model mean change is
greater than two standard deviations of internal variability and where at least 90% of models agree on the sign of change (see Box 12.1).
1073
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
changes in tropical atmospheric circulation are assessed in Section 2.7.5,
while Section 10.3.3 discusses attribution of these observed changes to
anthropogenic forcing. Evidence is inconclusive on recent trends in the
strength of the Hadley (Stachnik and Schumacher, 2011) and Walker
Circulations (Vecchi et al., 2006; Sohn and Park, 2010; Merrifield, 2011;
Luo et al., 2012; Tokinaga et al., 2012), though there is medium confi-
dence of an anthropogenic influence on the observed widening of the
Hadley Circulation (Hu and Fu, 2007; Johanson and Fu, 2009; Davis and
Rosenlof, 2012). In the projections, there are indications of a weakening
of tropical overturning of air as the climate warms (Held and Soden,
2006; Vecchi and Soden, 2007; Gastineau et al., 2008, 2009; Chou and
Chen, 2010; Chadwick et al., 2012; Bony et al., 2013). In the SRES A1B
scenario, CMIP3 models show a remarkable agreement in simulating a
weakening of the tropical atmospheric overturning circulation (Vecchi
and Soden, 2007). CMIP5 models also show a consistent weakening
(Chadwick et al., 2012). Along the ascending branches of tropical over-
turning cells, a reduction in convective mass flux from the boundary
layer to the free atmosphere is implied by the differential response to
global warming of the boundary-layer moisture content and surface
evaporation. This weakening of vertical motion along the ascending
regions of both the tropical meridional and near-equatorial zonal cells
is associated with an imbalance in the rate of atmospheric moisture
increase and that of global mean precipitation (Held and Soden, 2006).
A reduction in the compensating climatological subsidence along the
downward branches of overturning circulations, where the rate of
increase of static stability exceeds radiative cooling, is implied.
Several mechanisms have been suggested for the changes in the inten-
sity of the tropical overturning circulation. The weakening of low-level
convective mass flux along ascending regions of tropical overturning
cells has been ascribed to changes in the hydrologic cycle (Held and
Soden, 2006; Vecchi and Soden, 2007). Advection of dry air from sub-
sidence regions towards the ascending branches of large-scale tropical
circulation has been suggested to be a feasible mechanism weakening
ascent along the edges of convection regions (Chou et al., 2009). A
deepening of the tropical troposphere in response to global warming
increases the vertical extent of convection, which has been shown to
increase the atmosphere’s moist stability and thus also weakening
overturning cells (Chou and Chen, 2010). An imbalance between the
increase in diabatic heating of the troposphere and in static stabili-
ty whereby the latter increases more rapidly has also been thought
to play a role in weakening tropical ascent (Lu et al., 2008). Mean
advection of enhanced vertical stratification under GHG forcing which
involves cooling of convective regions and warming of subsidence
regions has been shown to slow down tropical cells (Ma et al., 2012).
The latest findings using CMIP5 models reveal that an increase in
GHGs ( particularly CO
2
) contributes significantly to weakening tropi-
cal overturning cells by reducing radiative cooling in the upper atmos-
phere (Bony et al., 2013). SST gradients have also been found to play
a role in altering the strength of tropical cells (Tokinaga et al., 2012;
Ma and Xie, 2013). Evidence has been provided suggesting that the SH
Hadley Cell may strengthen in response to meridional SST gradients
featuring reduced warming in the SH subtropical oceans relative to the
NH, particularly over the Pacific and Indian Oceans (Ma and Xie, 2013).
The north-to-south SST warming gradients are a source of intermodel
differences in their projections of changes in the SH Hadley Circulation.
Apart from changes in Hadley Circulation strength, a robust feature
in 21st century climate model simulations is an increase in the cell’s
depth and width (Mitas and Clement, 2006; Frierson et al., 2007; Lu
et al., 2007; Lu et al., 2008), with the latter change translating to a
broadening of tropical regions (Seidel and Randel, 2007; Seidel et al.,
2008) and a poleward displacement of subtropical dry zones (Lu et
al., 2007; Scheff and Frierson, 2012). The increase in the cell’s depth
is consistent with a tropical tropopause rise. The projected increase in
the height of the tropical tropopause and the associated increase in
meridional temperature gradients close to the tropopause slope have
been proposed to be an important mechanism behind the Hadley cell
expansion and the poleward displacement of the subtropical westerly
jet (Lu et al., 2008; Johanson and Fu, 2009). An increase in subtropical
and mid-latitude static stability has been found to be an important
factor widening the Hadley Cell by shifting baroclinic eddy activity and
the associated eddy-driven jet and subsidence poleward (Mitas and
Clement, 2006; Lu et al., 2008). The projected widening of the Hadley
Cell is consistent with late 20th century observations, where ~2° to 5°
latitude expansion was found (Fu et al., 2006; Johanson and Fu, 2009).
The consistency of simulated changes in CMIP3 and CMIP5 models and
the consistency of Hadley Cell changes with the projected tropopause
rise and increase in subtropical and mid-latitude static stability indi-
cate that a widening and weakening of the NH Hadley Cell by the late
21st century is likely.
The zonally asymmetric Walker Circulation is projected to weaken
under global warming (Power and Kociuba, 2011a, 2011b), more than
the Hadley Circulation (Lu et al., 2007; Vecchi and Soden, 2007). The
consistency of the projected Walker Circulation slowdown from CMIP3
to CMIP5 suggests that its change is robust (Ma and Xie, 2013). Almost
everywhere around the equatorial belt, changes in the 500 hPa ver-
tical motion oppose the climatological background motion, notably
over the maritime continent (Vecchi and Soden, 2007; Shongwe et al.,
2011). Around the Indo-Pacific warm pool, in response to a spatially
uniform SST warming, the climatological upper tropospheric diver-
gence weakens (Ma and Xie, 2013). Changes in the strength of the
Walker Circulation also appear to be linked to differential warming
between the Indian and Pacific Ocean warming at low latitudes (Luo et
al., 2012). Over the equatorial Pacific Ocean, where mid-tropospheric
ascent is projected to strengthen, changes in zonal SST and hence sea
level pressure gradients induce low-level westerly wind anomalies that
act to weaken the low-level branch of the Pacific Walker Circulation.
These projected changes in the tropical Pacific circulation are already
occurring (Zhang and Song, 2006). However, the projected weakening
of the Pacific Walker Cell does not imply an increase in the frequency
and/or magnitude of El Niño events (Collins et al., 2010). The consisten-
cy of simulated changes in CMIP3 and CMIP5 models and the consist-
ency of Walker Cell changes with equatorial SST and pressure-gradient
changes that are already observed indicate that a weakening of the
Walker Cell by the late 21st century is likely.
In the upper atmosphere, a robust feature of projected stratospheric
circulation change is that the Brewer–Dobson circulation will likely
strengthen in the 21st century (Butchart et al., 2006, 2010; Li et al.,
2008; McLandress and Shepherd, 2009; Shepherd and McLandress,
2011). In a majority of model experiments, the projected changes in
the large-scale overturning circulation in the stratosphere feature an
1074
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
intensification of tropical upward mass flux, which may extend to the
upper stratosphere. The proposed driver of the increase in mass flux at
the tropical lower stratosphere is the enhanced propagation of wave
activity, mainly resolved planetary waves, associated with a positive
trend in zonal wind structure (Butchart and Scaife, 2001; Garcia and
Randel, 2008). In the 21st century, increases in wave excitation from
diabatic heating in the upper tropical troposphere could reinforce the
wave forcing on the tropical upwelling branch of the stratospheric
mean meridional circulation (Calvo and Garcia, 2009). Parameterized
orographic gravity waves that result from strengthening of subtropical
westerly jets and cause more waves to propagate into the lower strat-
osphere also play a role (Sigmond et al., 2004; Butchart et al., 2006).
The projected intensification in tropical upwelling is counteracted by
enhanced mean extratropical/polar lower stratospheric subsidence. In
the NH high latitudes, the enhanced downwelling is associated with an
increase in stationary planetary wave activities (McLandress and Shep-
herd, 2009). The intensification of the stratospheric meridional residual
circulation has already been reported in studies focussing on the last
decades of the 20th century (Garcia and Randel, 2008; Li et al., 2008;
Young et al., 2012). The projected increase in troposphere-to-strato-
sphere mass exchange rate (Butchart et al., 2006) and stratospheric
mixing associated with the strengthening of the Brewer–Dobson circu-
lation will likely result in a decrease in the mean age of air in the lower
stratosphere. In the mid-latitude lower stratosphere, quasi-horizontal
mixing is a significant contributor to reducing the lifetimes of air. There
are some suggestions that the changes in stratospheric overturning
circulation could lead to a reduction in tropical ozone concentrations
and an increase at high latitudes (Jiang et al., 2007) and an increase
in the amplitude of the annual cycle of stratospheric ozone (Randel et
al., 2007).
12.4.4.3 Extratropical Storms: Tracks and Influences on
Planetary-Scale Circulation and Transports
Since the AR4, there has been continued evaluation of changes in
extratropical storm tracks under projected warming using both CMIP3
and, more recently, CMIP5 simulations, as well as supporting studies
using single models or idealized simulations. CMIP3 analyses use a
variety of methods for diagnosing storm tracks, but diagnosis of chang-
es in the tracks appears to be relatively insensitive to methods used
(Ulbrich et al., 2013). Analyses of SH storm tracks generally agree with
earlier studies, showing that extratropical storm tracks will tend to
shift poleward (Bengtsson et al., 2009; Gastineau et al., 2009; Gastin-
eau and Soden, 2009; Perrie et al., 2010; Schuenemann and Cassano,
2010; Chang et al., 2012b). The behaviour is consistent with a likely
trend in observed storm-track behaviour (see Section 2.7.6). Similar
behaviour appears in CMIP5 simulations for the SH (Figure 12.20c, d).
In SH winter there is a clear poleward shift in storm tracks of several
degrees and a reduction in storm frequency of only a few percent (not
shown). The poleward shift at the end of the century is consistent with
a poleward shift in the SH of the latitudes with strongest tropospheric
jets (Figure 12.19). This appears to coincide with shifts in baroclinic
dynamics governing extratropical storms (Frederiksen et al., 2011),
though the degree of jet shift appears to be sensitive to bias in a mod-
el’s contemporary-climate storm tracks (Chang et al., 2012a, 2012b).
Although there is thus some uncertainty in the degree of shift, the
consistency of behaviour with observation-based trends, consistency
between CMIP5 and CMIP3 projections under a variety of diagnostics
and the physical consistency of the storm response with other climatic
changes gives high confidence that a poleward shift of several degrees
in SH storm tracks is likely by the end of the 21st century under the
RCP8.5 scenario.
In the NH winter (Figure 12.20a, b), the CMIP5 multi-model ensemble
shows an overall reduced frequency of storms and less indication of
a poleward shift in the tracks. The clearest poleward shift in the NH
winter at the end of the 21st century occurs in the Asia-Pacific storm
track, where intensification of the westerly jet promotes more intense
cyclones in an ensemble of CMIP5 models (Mizuta, 2012). Otherwise,
changes in winter storm-track magnitude, as measured by band-pass
sea level pressure fluctuations, show only small change relative to
interannual and inter-decadal variability by the end of the 21st century
in SRES A1B and RCP4.5 simulations for several land areas over the NH
(Harvey et al., 2012). Consistency in CMIP3 and CMIP5 changes seen
in the SH are absent in the NH (Chang et al., 2012a). Factors identified
that affect changes in the North Atlantic basin’s storm track include
horizontal resolution (Colle et al., 2013) and how models simulate
changes in the Atlantic’s meridional overturning circulation (Catto et
al., 2011; Woollings et al., 2012), the zonal jet and Hadley Circulation
(Mizuta, 2012; Zappa et al., 2013) and subtropical upper troposphere
temperature (Haarsma et al., 2013). Substantial uncertainty and thus
low confidence remains in projecting changes in NH winter storm
tracks, especially for the North Atlantic basin.
Additional analyses of CMIP3 GCMs have determined other changes in
properties of extratropical storms. Most analyses find that the frequen-
cy of storms decreases in projected climates (Finnis et al., 2007; Favre
and Gershunov, 2009; Dowdy et al., 2013), though the occurrence of
strong storms may increase in some regions (Pinto et al., 2007; Bengts-
son et al., 2009; Ulbrich et al., 2009; Zappa et al., 2013). Many studies
focus on behaviour of specific regions, and results of these studies are
detailed in Section 14.6.2.
Changes in extratropical storms in turn may influence other large-scale
climatic changes. Kug et al. (2010) in a set of time-slice simulations
show that a poleward shift of storm tracks in the NH could enhance
polar warming and moistening. The Arctic Oscillation (AO) is sensitive
to synoptic eddy vorticity flux, so that projected changes in storm
tracks can alter the AO (Choi et al., 2010). The net result is that chang-
es in extratropical storms alter the climate in which they are embed-
ded, so that links between surface warming, extratropical storms and
their influence on climate are more complex than simple responses to
changes in baroclinicity (O’Gorman, 2010).
12.4.5 Changes in the Water Cycle
The water cycle consists of water stored on the Earth in all its phases,
along with the movement of water through the Earth’s climate system.
In the atmosphere, water occurs primarily as gaseous water vapour,
but it also occurs as solid ice and liquid water in clouds. The ocean is
primarily liquid water, but is partly covered by ice in polar regions. Ter-
restrial water in liquid form appears as surface water (lakes, rivers), soil
moisture and groundwater. Solid terrestrial water occurs in ice sheets,
glaciers, frozen lakes, snow and ice on the surface and permafrost.
1075
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
RCP4.5: 2081-2100
Southern Hemisphere JJANorthern Hemisphere DJF
RCP8.5: 2081-2100
b
cd
a
-3.9 -3.3 -2.7 -2.1 -1.5 -0.9 -0.3 0.3 0.9 1.5 2.1 2.7 3.3 3.9
(number density per month per unit area)
29 29
29 29
Figure 12.20 | Change in winter, extratropical storm track density (2081–2100) – (1986–2005) in CMIP5 multi-model ensembles: (a) RCP4.5 Northern Hemisphere December,
January and February (DJF) and (b) RCP8.5 Northern Hemisphere DJF, (c) RCP4.5 Southern Hemisphere June, July and August (JJA) and (d) RCP8.5 Southern Hemisphere JJA.
Storm-track computation uses the method of Bengtsson et al. (2006, their Figure 13a) applied to 6-hourly 850 hPa vorticity computed from horizontal winds in the CMIP5 archive.
The number of models used appears in the upper right of each panel. DJF panels include data for December 1985 and 2080 and exclude December 2005 and December 2100 for
in-season continuity. Stippling marks locations where at least 90% of the models agree on the sign of the change; note that this criterion differs from that used for many other
figures in this chapter, due to the small number of models providing sufficient data to estimate internal variability of 20-year means of storm-track statistics. Densities have units
(number density per month per unit area), where the unit area is equivalent to a 5° spherical cap (~10
6
km
2
). Locations where the scenario or contemporary-climate ensemble
average is below 0.5 density units are left white.
1076
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Projections of future changes in the water cycle are inextricably con-
nected to changes in the energy cycle (Section 12.4.3) and atmospheric
circulation (Section 12.4.4).
Saturation vapour pressure increases with temperature, but projected
future changes in the water cycle are far more complex than projected
temperature changes. Some regions of the world will be subject to
decreases in hydrologic activity while others will be subject to increas-
es. There are important local seasonal differences among the responses
of the water cycle to climate change as well.
At first sight, the water cycles simulated by CMIP3/5 models may
appear to be inconsistent, particularly at regional scales. Anthropogen-
ic changes to the water cycle are superimposed on complex naturally
varying modes of the climate (such as El Niño-Southern Oscillation
(ENSO), AO, Pacific Decadal Oscillation (PDO), etc.) aggravating the dif-
ferences between model projections. However, by careful consideration
of the interaction of the water cycle with changes in other aspects of
the climate system, the mechanisms of change are revealed, increasing
confidence in projections.
12.4.5.1 Atmospheric Humidity
Atmospheric water vapour is the primary GHG in the atmosphere. Its
changes affect all parts of the water cycle. However, the amount of
water vapour is dominated by naturally occurring processes and not
significantly affected directly by human activities. A common experi-
ence from past modelling studies is that relative humidity (RH) remains
approximately constant on climatological time scales and planetary
space scales, implying a strong constraint by the Clausius–Clapeyron
relationship on how specific humidity will change. The AR4 stated that
‘a broad-scale, quasi-unchanged RH response [to climate change] is
uncontroversial’ (Randall et al., 2007). However, underlying this fairly
straightforward behaviour are changes in RH that can influence chang-
es in cloud cover and atmospheric convection (Sherwood, 2010). More
recent analysis provides further detail and insight on RH changes. Anal-
ysis of CMIP3 and CMIP5 models shows near-surface RH decreasing
over most land areas as temperatures increase with the notable excep-
tion of parts of tropical Africa (O’Gorman and Muller, 2010) (Figure
12.21). The prime contributor to these decreases in RH over land is the
larger temperature increases over land than over ocean in the RCP sce-
narios (Joshi et al., 2008; Fasullo, 2010; O’Gorman and Muller, 2010).
The specific humidity of air originating over more slowly warming
oceans will be governed by saturation temperatures of oceanic air. As
this air moves over land and is warmed, its relative humidity drops as
any further moistening of the air over land is insufficient to maintain
constant RH, a behaviour Sherwood et al. (2010) term a last-satura-
tion-temperature constraint. The RH decrease over most land areas by
the end of the 21st century is consistent with a last-saturation-temper-
ature constraint and with observed behaviour during the first decade
of the current century (Section 2.5.5; Simmons et al., 2010). Land–
ocean differences in warming are projected to continue through the
21st century, and although the CMIP5 projected changes are small,
they are consistent with a last-saturation constraint, indicating with
medium confidence that reductions in near-surface RH over many land
areas are likely.
12.4.5.2 Patterns of Projected Average Precipitation Changes
Global mean precipitation changes have been presented in Section
12.4.1.1. The processes that govern large-scale changes in precipita-
tion are presented in Section 7.6, and are used here to interpret the
Figure 12.21 | Projected changes in near-surface relative humidity from the CMIP5 models under RCP8.5 for the December, January and February (DJF, left), June, July and August
(JJA, middle) and annual mean (ANN, right) averages relative to 1986–2005 for the periods 2046–2065 (top row), 2081–2100 (bottom row). The changes are differences in relative
humidity percentage (as opposed to a fractional or relative change). Hatching indicates regions where the multi-model mean change is less than one standard deviation of internal
variability. Stippling indicates regions where the multi-model mean change is greater than two standard deviations of internal variability and where at least 90% of models agree
on the sign of change (see Box 12.1).
1077
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
projected changes in RCP scenarios. Changes in precipitation extremes
are presented in Section 12.4.5.5. Further discussion of regional chang-
es, in particular the monsoon systems, is presented in Chapter 14.
Figure 12.22 shows the CMIP5 multi-model average percentage
change in seasonal mean precipitation in the middle of the 21st
century, at the end of the 21st century and at the end of the 22nd
century for the RCP8.5 scenario relative to the 1986–2005 average.
Precipitation changes for all the scenarios are shown in Annex I Sup-
plementary Material and scale approximately with the global mean
temperature (Section 12.4.2). In many regions, changes in precipitation
exhibit strong seasonal characteristics so that, in regions where the
sign of the precipitation changes varies with the season, the annual
mean values (Figure 12.10) may hide some of these seasonal changes,
resulting in weaker confidence than seasonal mean values (Chou et al.,
2013; Huang et al., 2013).
The patterns of multi-model precipitation changes displayed in Figure
12.22 tend to smooth and decrease the spatial contrast of precip-
itation changes simulated by each model, in particular over regions
where model results disagree. Thus the amplitude of the multi-model
ensemble mean precipitation response significantly underestimates
the median amplitude computed from each individual model (Neelin
et al., 2006; Knutti et al., 2010a). The CMIP3/5 multi-model ensemble
precipitation projections must be interpreted in the context of uncer-
tainty. Multi-model projections are not probabilistic statements about
the likelihood of changes. Maps of multi-model projected changes are
smoothly varying but observed changes are and will continue to be
much more granular.
To analyze the patterns of projected precipitation changes, a useful
framework consists in decomposing them into a part that is related to
atmospheric circulation changes and a part that is related mostly to
water vapour changes, referred to as dynamical and thermodynamical
components, respectively. However, the definition of these two com-
ponents may differ among studies. At the time of the AR4, the robust
changes of the difference between precipitation and evaporation
(P E) were interpreted as a wet-get-wetter and dry-get-drier type
of response (Mitchell et al., 1987; Chou and Neelin, 2004; Held and
Soden, 2006). The theoretical background, which is more relevant over
oceans than over land, is that the lower-tropospheric water vapour
increase with temperature enhances the moisture transported by
the circulation. This leads to additional moisture convergence within
the convergence zones and to additional moisture divergence in the
descent zones, increasing the contrast in precipitation minus evapo-
ration values between moisture convergence and divergence regions.
A weakening of the tropical overturning circulation (see Section
12.4.4.2) partially opposes this thermodynamic response (Chou and
Neelin, 2004; Held and Soden, 2006; Vecchi and Soden, 2007; Chou
et al., 2009; Seager et al., 2010; Allan, 2012; Bony et al., 2013). At the
regional scale the dynamic response may be larger than the thermo-
dynamic response, and this has been analyzed in more detail since
the AR4 (Chou et al., 2009; Seager et al., 2010; Xie et al., 2010; Muller
and O’Gorman, 2011; Chadwick et al., 2012; Scheff and Frierson, 2012;
Bony et al., 2013; Ma and Xie, 2013). Over continents, this simple wet-
get-wetter and dry-get-drier type of response fails for some important
regions such as the Amazon. At the global scale, the net water vapour
transport from oceans to land increases, and therefore the average P
E over continents also increases (Liepert and Previdi, 2012).
In the mid and high latitudes, a common feature across generations of
climate models is a simulated increased precipitation. The thermody-
namical component explains most of the projected increase (Emori and
Brown, 2005; Seager et al., 2010). This is consistent with theoretical
explanations assuming fixed atmospheric flow patterns but increased
water vapour in the lower troposphere (Held and Soden, 2006). In addi-
tion to this thermodynamical effect, water transport may be modified
by the poleward shift of the storm tracks and by the increase of their
intensity (Seager et al., 2010; Wu et al., 2011b), although confidence in
such changes in storm tracks may not be high (see Section 12.4.4). On
seasonal time scales, the minimum and maximum values of precipita-
tion both increase, with a larger increase of the maximum and there-
fore an increase of the annual precipitation range (Seager et al., 2010;
Chou and Lan, 2012). In particular, the largest changes over northern
Eurasia and North America are projected to occur during winter. At
high latitudes of the NH, the precipitation increase may lead to an
increase of snowfall in the colder regions and a decrease of snowfall
in the warmer regions due to the decreased number of freezing days
(see Section 12.4.6.2).
Most models simulate a large increase of the annual mean precipita-
tion over the equatorial ocean and an equatorward shift of the Inter-
tropical Convergence Zone (ITCZ), in both summer and winter seasons,
that are mainly explained by atmospheric circulation changes (Chou et
al., 2009; Seager et al., 2010; Sobel and Camargo, 2011). The chang-
es of the atmospheric circulation have different origins. Along the
margins of the convection zones, spatial inhomogeneities, including
local convergence feedback or the rate at which air masses from dry
regions tend to flow into the convection zone, can yield a considerable
sensitivity in precipitation response (Chou et al., 2006; Neelin et al.,
2006). Along the equator, atmosphere–ocean interactions yield to a
maximum of SST warming and a large precipitation increase there (Xie
et al., 2010; Ma and Xie, 2013). Model studies with idealized configu-
rations suggest that tropical precipitation changes should be interpret-
ed as responses to changes of the atmospheric energy budget rather
than responses to changes of SST (Kang and Held, 2012). All of these
atmospheric circulation changes, and therefore precipitation changes,
can differ considerably from model to model. This is the case over both
ocean and land. For instance, the spread of model projections in the
Sahel region, West Africa, is large in both the CMIP3 and CMIP5 mul-
ti-model data base (Roehrig et al., 2013).
In the subtropical dry regions, there is a robust decrease of PE that
is accounted for by the thermodynamic contribution (Chou and Neelin,
2004; Held and Soden, 2006; Chou et al., 2009; Seager et al., 2010;
Bony et al., 2013). Over ocean, the spatial heterogeneity of temperature
increase impacts the lower-tropospheric water vapour increase, which
impacts both the thermodynamic and the dynamic responses (Xie et
al., 2010; Ma and Xie, 2013). In addition, the pattern of precipitation
changes in dry regions may be different from that of PE because the
contribution of evaporation changes can be as large (but of opposite
sign) as the moisture transport changes (Chou and Lan, 2012; Scheff
and Frierson, 2012; Bony et al., 2013). This is especially the case over
the subsidence regions during the warm season over land where the
1078
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
agreement between models is the smallest (Chou et al., 2009; Allan,
2012). A robust feature is the decline of precipitation on the poleward
flanks of the subtropical dry zones as a consequence of the Hadley Cell
expansion, with possible additional decrease from a poleward shift of
the mid latitude storm tracks (Seager et al., 2010; Scheff and Frierson,
2012). On seasonal time scales, the minimum and the maximum values
of precipitation both increase, with a larger increase of the maximum
and therefore an increase of the annual precipitation range (Sobel and
Camargo, 2011; Chou and Lan, 2012).
Long-term precipitation changes are driven mainly by the increase of
the surface temperature, as presented above, but other factors also
contribute to them. Recent studies suggest that CO
2
increase has a sig-
nificant direct influence on atmospheric circulation, and therefore on
global and tropical precipitation changes (Andrews et al., 2010; Bala et
al., 2010; Cao et al., 2012; Bony et al., 2013). Over the ocean, the pos-
itive RF from increased atmospheric CO
2
reduces the radiative cooling
of the troposphere and the large scale rising motion and hence reduc-
es precipitation in the convective regions. Over large landmasses, the
direct effect of CO
2
on precipitation is the opposite owing to the small
thermal inertia of land surfaces (Andrews et al., 2010; Bala et al., 2010;
Cao et al., 2012; Bony et al., 2013). Regional precipitation changes are
also influenced by aerosol and ozone (Ramanathan et al., 2001; Allen
et al., 2012; Shindell et al., 2013a) through both local and large-scale
processes, including changes in the circulation. Stratospheric ozone
depletion contributes to the poleward expansion of the Hadley Cell
and the related change of precipitation in the SH (Kang et al., 2011)
whereas black carbon and tropospheric ozone increases are major con-
tributors in the NH (Allen et al., 2012). Regional precipitation changes
depend on regional forcings and on how models simulate their local
and remote effects. Based on CMIP3 results, the inter-model spread
of the estimate of precipitation changes over land is larger than the
inter-scenario spread except in East Asia (Frieler et al., 2012).
Seasonal mean percentage precipitation change (RCP8.5)
Figure 12.22 | Multi-model CMIP5 average percentage change in seasonal mean precipitation relative to the reference period 1986–2005 averaged over the periods 2045–2065,
2081–2100 and 2181–2200 under the RCP8.5 forcing scenario. Hatching indicates regions where the multi-model mean change is less than one standard deviation of internal
variability. Stippling indicates regions where the multi-model mean change is greater than two standard deviations of internal variability and where at least 90% of models agree
on the sign of change (see Box 12.1).
1079
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Projected precipitation changes vary greatly between models, much
more so than for temperature projections. Part of this variance is due to
genuine differences between the models including their ability to rep-
licate observed precipitation patterns (see Section 9.4.1.1). However, a
large part of it is also the result of the small ensemble size from each
model (Rowell, 2012). This is especially true for regions of small pro-
jected changes located between two regions: one experiencing signif-
icant increases while the other experiences significant decreases. Indi-
vidual climate model realizations will differ in their projection of future
precipitation changes in these regions simply owing to their internal
variability (Deser et al., 2012b; Deser et al., 2012a). Multi-model pro-
jections containing large numbers of realizations would tend to feature
small changes in these regions, and hatching in Figure 12.22 indicates
regions where the projected multi-model mean change is less than one
standard deviation of internal variability (method (a), Box 12.1). Confi-
dence in projections in regions of limited or no change in precipitation
may be more difficult to obtain than confidence in regions of large pro-
jected changes. However, Power et al. (2012) and Tebaldi et al. (2011)
show that for some of the regions featuring small multi-model average
projected changes, effective consensus in projections may be better
than the metrics reported in AR4 would imply.
Since the AR4, progress has been made in the understanding of the
processes that control large scale precipitation changes. There is high
confidence that the contrast of seasonal mean precipitation between
dry and wet regions will increase in a warmer climate over most of
the globe although there may be regional exceptions to this general
pattern. This response is particularly robust when considering PE
changes as a function of atmospheric dynamical regimes. However, it
is important to note that significant exceptions can occur in specific
regions especially along the equator and on the poleward edges of the
subtropical dry zone. In these regions, atmospheric circulation changes
lead to shifts of the precipitation patterns. There is high confidence that
the contrast between wet and dry seasons will increase over most of
the globe as temperatures increase. Over the mid- and high-latitude
regions, projected precipitation increases in winter are larger than in
summer. Over most of the subtropical oceans, projected precipitation
increases in summer are larger than in winter.
The changes in precipitation shown in Figure 12.22 exhibit patterns
that become more pronounced and confidence in them increases
as temperatures increase. More generally, the spatial and temporal
changes in precipitation between two scenarios or within two peri-
ods of a given scenario exhibit the pattern scaling behavior and lim-
itations described in Section 12.4.2. The patterns and the associated
multi-model spreads in CMIP5 for the RCP scenarios are very similar
to those in CMIP3 for the SRES scenarios discussed in the AR4, with
the projections in CMIP5 being slightly more consistent over land than
those from CMIP3 (Knutti and Sedláček, 2013). The largest percentage
changes are at the high latitudes. By the end of the 21st century, over
the large northern land masses, increased precipitation is likely under
the RCP8.5 scenario in the winter and spring poleward of 50°N. The
robustness across scenarios, the magnitude of the projected changes
versus natural variability and physical explanations described above
yield high confidence that the projected changes would be larger than
natural 20-year variations (see Box 12.1). In the tropics, precipitation
changes exhibit strong regional contrasts, with increased precipitation
over the equatorial Pacific and Indian Oceans and decreases over much
of the subtropical ocean. However, decreases are not projected to be
larger than natural 20-year variations anywhere until the end of this
century under the RCP8.5 scenario. Decreased precipitation in the
Mediterranean, Caribbean and Central America, southwestern United
States and South Africa is likely under the RCP8.5 scenario and is pro-
jected with medium confidence to be larger than natural variations by
the end of the 22nd century in some seasons (Box 12.1). The CMIP3
models’ historical simulations of zonal mean precipitation trends were
shown to underestimate observed trends (Gillett et al., 2004; Lambert
et al., 2005; Zhang et al., 2007; Liepert and Previdi, 2009) (see Section
10.3.2.2). Therefore it is more likely than not that the magnitude of the
projected future changes in Figure 12.22 based on the multi-model
mean is underestimated. Observational uncertainties including limited
global coverage and large natural variability, in addition to challenges
in precipitation modelling, limit confidence in assessment of climatic
changes in precipitation.
12.4.5.3 Soil Moisture
Near-surface soil moisture is the net result of a suite of complex process-
es (e.g., precipitation evapotranspiration, drainage, overland flow, infil-
tration), and heterogeneous and difficult-to-characterize aboveground
and belowground system properties (e.g., slope, soil texture). As a
result, regional to global-scale simulations of soil moisture and drought
remain relatively uncertain (Burke and Brown, 2008; Henderson-Sellers
et al., 2008). The AR4 (Section 8.2.3.2) discussed the lack of assess-
ments of global-scale models in their ability to simulate soil moisture,
and this problem appears to have persisted (Section 9.4.4.2). Further-
more, consistent multi-model projections of total soil moisture are diffi-
cult to make owing to substantial differences between climate models
in the depth of their soil. However, Koster et al. (2009a) argued that
once climatological statistics affecting soil moisture were accounted for,
different models tend to agree on soil moisture projections.
The AR4 summarized multi-model projections of 21st century annual
mean soil moisture changes as decreasing in the subtropics and Med-
iterranean region, and increasing in east Africa and central Asia. Dai
(2013) found similar changes in an ensemble of 11 CMIP5 GCMs under
RCP4.5. Figure 12.23 shows projected changes in surface soil moisture
(upper 10 cm) in the CMIP5 ensemble at the end of the 21st century
under the RCPs 2.6, 4.5, 6.0 and 8.5. We focus on this new CMIP5
specification because it describes soil moisture at a consistent depth
across all CMIP5 models. The broad patterns are moderately consist-
ent across the RCPs, with the changes tending to become stronger as
the strength of the forcing change increases. The agreement among
CMIP5 models and the consistency with other physical features of
climate change indicate high confidence in certain regions where
surface soils are projected to dry. There is little-to-no confidence any-
where in projections of moister surface soils. Under RCP8.5, with the
largest projected change, individual ensemble members (not shown)
show consistency across the ensemble for drying in the Mediterranean
region, northeast and southwest South America, southern Africa, and
southwestern USA. However, ensemble members show disagreement
on the sign of change in large regions such as central Asia or the high
northern latitudes. The Mediterranean, southwestern USA, northeast
South America and southern African drying regions are consistent with
1080
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
projected widening of the Hadley Circulation that shifts downwelling,
thus inhibiting precipitation in these regions. The large-scale drying in
the Mediterranean, southwest USA, and southern Africa appear across
generations of projections and climate models and is deemed likely
as global temperatures rise and will increase the risk of agricultural
drought. In addition, an analysis of CMIP3 and CMIP5 projections of
soil moisture in five drought-prone regions indicates that the differ-
ences in future forcing scenarios are the largest source of uncertain-
ty in such regions rather than differences between model responses
(Orlowsky and Seneviratne, 2012).
Other recent assessments include multi-model ensemble approaches,
dynamical downscaling, and regional climate models applied around
the globe and illustrate the variety of issues influencing soil moisture
changes. Analyses of the southwestern USA using CMIP3 models
(Christensen and Lettenmaier, 2007; Seager et al., 2007) show consist-
ent projections of drying, primarily due to a decrease in winter precipi-
tation. In contrast, Kellomaki et al. (2010) find that SRES A2 projections
for Finland yield decreased snow depth, but soil moisture generally
increases, consistent with the general increase in precipitation occur-
ring in high northern latitudes. Kolomyts and Surova (2010), using pro-
jections from the CMIP3 models, GISS and HadCM2, under the SRES
A2 forcing, show that vegetation type has substantial influence on the
development of pronounced drying over the 21st century in Middle
Volga Region forests.
Projected changes in soil moisture from the CMIP3/5 models also show
substantial seasonal variation. For example, soil moisture changes in
the North American midlatitudes, coupled with projected warming,
increases the strength of land–atmosphere coupling during spring and
summer in 15 GCMs under RCP8.5 (Dirmeyer et al., 2013). For the
Cline River watershed in western Canada, Kienzle et al. (2012) find
decreases in summer soil moisture content, but annual increases aver-
aging 2.6% by the 2080s using a suite of CMIP3 GCMs simulating B1,
A1B and A2 scenarios to drive a regional hydrology model. Hansen et
al. (2007), using dynamical downscaling of one GCM running the A2
scenario, find summer soil moisture decreases in Mongolia of up to
6% due to increased potential evaporation in a warming climate and
decreased precipitation and decreased precipitation.
Soil moisture projections in high latitude permafrost regions are crit-
ically important for assessing future climate feedbacks from trace-
gas emissions (Zhuang et al., 2004; Riley et al., 2011) and vegetation
changes (Chapin et al., 2005). In addition to changes in precipitation,
snow cover and evapotranspiration, future changes in high-latitude
soil moisture also will depend on permafrost degradation, thermokarst
evolution, rapid changes in drainage (Smith et al., 2005), and changes
in plant communities and their water demands. Current understanding
of these interacting processes at scales relevant to climate is poor, so
that full incorporation in current GCMs is lacking.
Figure 12.23 | Change in annual mean soil moisture (mass of water in all phases in the uppermost 10 cm of the soil) (mm) relative to the reference period 1986–2005 projected
for 2081–2100 from the CMIP5 ensemble. Hatching indicates regions where the multi-model mean change is less than one standard deviation of internal variability. Stippling
indicates regions where the multi-model mean change is greater than two standard deviations of internal variability and where at least 90% of models agree on the sign of change
(see Box 12.1). The number of CMIP5 models used is indicated in the upper right corner of each panel.
Annual mean near-surface soil moisture change (2081-2100)
1081
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
12.4.5.4 Runoff and Evaporation
In the AR4, 21st century model-projected runoff consistently showed
decreases in southern Europe, the Middle East, and southwestern USA
and increases in Southeast Asia, tropical East Africa and at high north-
ern latitudes. The same general features appear in the CMIP5 ensemble
of GCMs for all four RCPs shown in Figure 12.24, with the areas of most
robust change typically increasing with magnitude of forcing change.
However, the robustness of runoff decreases in the southwestern USA
is less in the CMIP5 models compared to the AR4. The large decreases
in runoff in southern Europe and southern Africa are consistent with
changes in the Hadley Circulation and related precipitation decreases
and warming-induced evapotranspiration increases. The high northern
latitude runoff increases are likely under RCP8.5 and consistent with
the projected precipitation increases (Figure 12.22). The consistency of
changes across different generations of models and different forcing
scenarios, together with the physical consistency of change indicates
that decreases are also likely in runoff in southern Europe, the Middle
East, and southern Africa in this scenario.
A number of reports since the AR4 have updated findings from CMIP3
models and analyzed a large set of mechanisms affecting runoff. Sev-
eral studies have focussed on the Colorado River basin in the United
States (Christensen and Lettenmaier, 2007; McCabe and Wolock, 2007;
Barnett and Pierce, 2008; Barnett et al., 2008) showing that runoff
reductions that do happen under global warming occur through a
combination of evapotranspiration increases and precipitation decreas-
es, with the overall reduction in river flow exacerbated by human water
demands on the basin’s supply.
A number of CMIP3 analyses have examined trends and seasonal shifts
in runoff. For example, Kienzle et al. (2012) studied climate change sce-
narios over the Cline River watershed in western Canada and projected
(1) spring runoff and peak streamflow up to 4 weeks earlier than in
1961–1990; (2) significantly higher streamflow between October and
June; and (3) lower streamflow between July and September. For the
Mediterranean basin, an ensemble of regional climate models driven
by several GCMs using the A1B scenario have a robust decrease in
runoff emerging only after 2050 (Sanchez-Gomez et al., 2009).
Annual mean surface evaporation in the models assessed in AR4
showed increases over most of the ocean and increases or decreases
over land with largely the same pattern over land as increases and
decreases in precipitation. Similar behaviour occurs in an ensemble of
CMIP5 models (Figure 12.25). Evaporation increases over most of the
ocean and land, with prominent areas of decrease over land occurring
in southern Africa and northwestern Africa along the Mediterranean.
The areas of decrease correspond to areas with reduced precipitation.
There is some uncertainty about storm-track changes over Europe (see
Sections 12.4.3 and 14.6.2). However, the consistency of the decreas-
es across different generations of models and different forcing sce-
narios along with the physical basis for the precipitation decrease
Figure 12.24 | Change in annual mean runoff relative to the reference period 1986–2005 projected for 2081–2100 from the CMIP5 ensemble. Hatching indicates regions where
the multi-model mean change is less than one standard deviation of internal variability. Stippling indicates regions where the multi-model mean change is greater than two standard
deviations of internal variability and where at least 90% of models agree on the sign of change (see Box 12.1). The number of CMIP5 models used is indicated in the upper right
corner of each panel.
1082
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
indicates that these decreases in annual mean evaporation are likely
under RCP8.5, but with medium confidence. Annual mean evapora-
tion increases over land in the northern high latitudes are consistent
with the increase in precipitation and the overall warming that would
increase potential evaporation. For the northern high latitudes, the
physical consistency and the similar behaviour across multiple gener-
ations and forcing scenarios indicates that annual mean evaporation
increases there are likely, with high confidence.
Evapotranspiration changes partly reflect changes in precipitation.
However, some changes might come from altered biological processes.
For example, increased atmospheric CO
2
promotes stomatal closure
and reduced transpiration (Betts et al., 2007; Cruz et al., 2010) which
can potentially yield increased runoff. There is potential for substan-
tial feedback between vegetation changes and regional water cycles,
though the impact of such feedback remains uncertain at this point
due to limitations on modelling crop and other vegetation processes in
GCMs (e.g., Newlands et al., 2012) and uncertainties in plant response,
ecosystem shifts and land management changes.
12.4.5.5 Extreme Events in the Water Cycle
In addition to the changes in the seasonal pattern of mean precipitation
described above, the distribution of precipitation events is projected to
undergo profound changes (Gutowski et al., 2007; Sun et al., 2007;
Boberg et al., 2010). At daily to weekly scales, a shift to more intense
individual storms and fewer weak storms is projected (Seneviratne et
al., 2012). At seasonal or longer time scales, increased evapotranspira-
tion over land can lead to more frequent and more intense periods of
agricultural drought.
A general relationship between changes in total precipitation and
extreme precipitation does not exist (Seneviratne et al., 2012). Two
possible mechanisms controlling short-term extreme precipitation
amounts are discussed at length in the literature and are similar to the
thermodynamic and dynamical mechanisms detailed above for chang-
es in average precipitation.
The first considers that extreme precipitation events occur when most
of the available atmospheric water vapour rapidly precipitates out in a
single storm. The maximum amount of water vapour in air (saturation)
is determined by the Clausius–Clapeyron relationship. As air temper-
ature increases, this saturated amount of water also increases (Allen
and Ingram, 2002; Pall et al., 2007; Allan and Soden, 2008; Kendon et
al., 2010). Kunkel et al. (2013) examined the CMIP5 model RCP4.5 and
8.5 projections for changes in maximum water vapour concentrations,
a principal factor controlling the probable bound on maximum precipi-
tation, concluding that maximum water vapour changes are compara-
ble to mean water vapour changes but that the potential for changes
in dynamical factors is less compelling. Such increases in atmospheric
water vapour are expected to increase the intensity of individual pre-
cipitation events, but have less impact on their frequency. As a result
Figure 12.25 | Change in annual mean evaporation relative to the reference period 1986–2005 projected for 2081–2100 from the CMIP5 ensemble. Hatching indicates regions
where the multi-model mean change is less than one standard deviation of internal variability. Stippling indicates regions where the multi-model mean change is greater than two
standard deviations of internal variability and where at least 90% of models agree on the sign of change (see Box 12.1). The number of CMIP5 models used is indicated in the
upper right corner of each panel.
1083
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
projected increases in extreme precipitation may be more reliable than
similar projections of changes in mean precipitation in some regions
(Kendon et al., 2010).
A second mechanism for extreme precipitation put forth by O’Gorman
and Schneider (2009a, 2009b) is that such events are controlled by
anomalous horizontal moisture flux convergence and associated con-
vective updrafts which would change in a more complicated fashion
in a warmer world (Sugiyama et al., 2010). Emori and Brown (2005)
showed that the thermodynamic mechanism dominated over the
dynamical mechanism nearly everywhere outside the tropical warm
pool. However, Utsumi et al. (2011) used gridded observed daily data
to find that daily extreme precipitation monotonically increases with
temperature only at high latitudes, with the opposite behaviour in
the tropics and a mix in the mid-latitudes. Li et al. (2011a) found that
both mechanisms contribute to extreme precipitation in a high-res-
olution aquaplanet model with updrafts as the controlling element
in the tropics and air temperature controlling the mid-latitudes con-
sistent with the results by Chou et al. (2012). Using a high-resolution
regional model, Berg et al. (2009) found a seasonal dependence in
Europe with the Clausius–Clapeyron relationship providing an upper
limit to daily precipitation intensity in winter but water availability
rather than storage capacity is the controlling factor in summer. Addi-
tionally, Lenderink and Van Meijgaard (2008) found that very short
(sub-daily) extreme precipitation events increase at a rate twice the
amount predicted by Clausius–Clapeyron scaling in a very high-resolu-
tion model over Europe suggesting that both mechanisms can interact
jointly. Gastineau and Soden (2009) found in the CMIP3 models that
the updrafts associated with the most extreme tropical precipitation
events actually weaken despite an increase in the frequency of the
heaviest rain rates further complicating simple mechanistic explana-
tions. See also Sections 7.6.5 and 11.3.2.5.2.
Projections of changes in future extreme precipitation may be larger
at the regional scales than for future mean precipitation, but natural
variability is also larger causing a tendency for signal-to-noise ratios
to decrease when considering increasingly extreme metrics. However,
mechanisms of natural variability still are a large factor in assessing
the robustness of projections (Kendon et al., 2008). In addition, large-
scale circulation changes, which are uncertain, could dominate over the
above mechanisms depending on the rarity and type of events consid-
ered. However, analysis of CMIP3 models suggests circulation changes
are potentially insufficient to offset the influence of increasing atmos-
pheric water vapour on extreme precipitation change over Europe at
least on large spatial scales (Kendon et al., 2010). An additional shift of
the storm track has been shown in models with a better representation
of the stratosphere, and this is found to lead to an enhanced increase
in extreme rainfall over Europe in winter (Scaife et al., 2012).
Similar to temperature extremes (Section 12.4.3.3), the definition of
a precipitation extreme depends very much on context and is often
used in discussion of particular climate-related impacts (Seneviratne
et al. (2012), Box 3.1). Consistently, climate models project future epi-
sodes of more intense precipitation in the wet seasons for most of the
land areas, especially in the NH and its higher latitudes, and the mon-
soon regions of the world, and at a global average scale. The actual
magnitude of the projected change is dependent on the model used,
1960 1980 2000 2020 2040 2060 2080 2100
−5
0
5
10
15
20
Year
−5
0
5
10
15
20
Wettest consecutive five days (RX5day)
Relative change (%)
a)
historical
RCP2.6
RCP4.5
RCP8.5
CMIP3 B1 CMIP3 A1B CMIP3 A2
18
18
Figure 12.26 | (a, b) Projected percent changes (relative to the 1981–2000 refer-
ence period in common with CMIP3) from the CMIP5 models in RX5day, the annual
maximum five-day precipitation accumulation. (a) Global average percent change over
land regions for the RCP2.6, RCP4.5 and RCP8.5 scenarios. Shading in the time series
represents the interquartile ensemble spread (25th and 75th quantiles). The box-and-
whisker plots show the interquartile ensemble spread (box) and outliers (whiskers) for
11 CMIP3 model simulations of the SRES scenarios A2 (orange), A1B (cyan) and B1
(purple) globally averaged over the respective future time periods (2046–2065 and
2081–2100) as anomalies from the 1981–2000 reference period. (b) Percent change
over the 2081–2100 period in the RCP8.5 scenario. (c) Projected change in annual
CDD, the maximum number of consecutive dry days when precipitation is less than 1
mm, over the 2081–2100 period in the RCP8.5 scenario (relative to the 1981–2000
reference period) from the CMIP5 models. Stippling indicates gridpoints with changes
that are significant at the 5% level using a Wilcoxon signed-ranked test. (Updated from
Sillmann et al. (2013), excluding the FGOALS-s2 model.)
1084
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Frequently Asked Questions
FAQ 12.2 | How Will the Earth’s Water Cycle Change?
The flow and storage of water in the Earth’s climate system are highly variable, but changes beyond those due to
natural variability are expected by the end of the current century. In a warmer world, there will be net increases in
rainfall, surface evaporation and plant transpiration. However, there will be substantial differences in the changes
between locations. Some places will experience more precipitation and an accumulation of water on land. In others,
the amount of water will decrease, due to regional drying and loss of snow and ice cover.
The water cycle consists of water stored on the Earth in all its phases, along with the movement of water through
the Earth’s climate system. In the atmosphere, water occurs primarily as a gas—water vapour—but it also occurs as
ice and liquid water in clouds. The ocean, of course, is primarily liquid water, but the ocean is also partly covered by
ice in polar regions. Terrestrial water in liquid form appears as surface water—such as lakes and rivers—soil moisture
and groundwater. Solid terrestrial water occurs in ice sheets, glaciers, snow and ice on the surface and in permafrost
and seasonally frozen soil.
Statements about future climate sometimes say that the water cycle will accelerate, but this can be misleading, for
strictly speaking, it implies that the cycling of water will occur more and more quickly with time and at all locations.
Parts of the world will indeed experience intensification of the water cycle, with larger transports of water and
more rapid movement of water into and out of storage reservoirs. However, other parts of the climate system will
experience substantial depletion of water, and thus less movement of water. Some stores of water may even vanish.
As the Earth warms, some general features of change will occur simply in response to a warmer climate. Those
changes are governed by the amount of energy that global warming adds to the climate system. Ice in all forms will
melt more rapidly, and be less pervasive. For example, for some simulations assessed in this report, summer Arctic
sea ice disappears before the middle of this century. The atmosphere will have more water vapour, and observations
and model results indicate that it already does. By the end of the 21st century, the average amount of water vapour
in the atmosphere could increase by 5 to 25%, depending on the amount of human emissions of greenhouse gases
and radiatively active particles, such as smoke. Water will evaporate more quickly from the surface. Sea level will
rise due to expansion of warming ocean waters and melting land ice flowing into the ocean (see FAQ 13.2).
These general changes are modified by the complexity of the climate system, so that they should not be expected
to occur equally in all locations or at the same pace. For example, circulation of water in the atmosphere, on land
and in the ocean can change as climate changes, concentrating water in some locations and depleting it in others.
The changes also may vary throughout the year: some seasons tend to be wetter than others. Thus, model simu-
lations assessed in this report show that winter precipitation in northern Asia may increase by more than 50%,
whereas summer precipitation there is projected to hardly change. Humans also intervene directly in the water
cycle, through water management and changes in land use. Changing population distributions and water practices
would produce further changes in the water cycle.
Water cycle processes can occur over minutes, hours, days and longer, and over distances from metres to kilometres
and greater. Variability on these scales is typically greater than for temperature, so climate changes in precipitation
are harder to discern. Despite this complexity, projections of future climate show changes that are common across
many models and climate forcing scenarios. Similar changes were reported in the AR4. These results collectively
suggest well understood mechanisms of change, even if magnitudes vary with model and forcing. We focus here
on changes over land, where changes in the water cycle have their largest impact on human and natural systems.
Projected climate changes from simulations assessed in this report (shown schematically in FAQ 12.2, Figure 1) gen-
erally show an increase in precipitation in parts of the deep tropics and polar latitudes that could exceed 50% by the
end of the 21st century under the most extreme emissions scenario. In contrast, large areas of the subtropics could
have decreases of 30% or more. In the tropics, these changes appear to be governed by increases in atmospheric
water vapour and changes in atmospheric circulation that further concentrate water vapour in the tropics and thus
promote more tropical rainfall. In the subtropics, these circulation changes simultaneously promote less rainfall
despite warming in these regions. Because the subtropics are home to most of the world’s deserts, these changes
imply increasing aridity in already dry areas, and possible expansion of deserts. (continued on next page)
1085
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
FAQ 12.2 (continued)
Increases at higher latitudes are governed by warmer temperatures, which allow more water in the atmosphere and
thus, more water that can precipitate. The warmer climate also allows storm systems in the extratropics to transport
more water vapour into the higher latitudes, without requiring substantial changes in typical wind strength. As
indicated above, high latitude changes are more pronounced during the colder seasons.
Whether land becomes drier or wetter depends partly on precipitation changes, but also on changes in surface
evaporation and transpiration from plants (together called evapotranspiration). Because a warmer atmosphere
can have more water vapour, it can induce greater evapotranspiration, given sufficient terrestrial water. However,
increased carbon dioxide in the atmosphere reduces a plant’s tendency to transpire into the atmosphere, partly
counteracting the effect of warming.
In the tropics, increased evapotranspiration tends to counteract the effects of increased precipitation on soil mois-
ture, whereas in the subtropics, already low amounts of soil moisture allow for little change in evapotranspiration.
At higher latitudes, the increased precipitation generally outweighs increased evapotranspiration in projected cli-
mates, yielding increased annual mean runoff, but mixed changes in soil moisture. As implied by circulation changes
in FAQ 12.2, Figure 1, boundaries of high or low moisture regions may also shift.
A further complicating factor is the character of rainfall. Model projections show rainfall becoming more intense,
in part because more moisture will be present in the atmosphere. Thus, for simulations assessed in this report, over
much of the land, 1-day precipitation events that currently occur on average every 20 years could occur every 10
years or even more frequently by the end of the 21st century. At the same time, projections also show that precipi-
tation events overall will tend to occur less frequently.
These changes produce two seemingly contradictory
effects: more intense downpours, leading to more
floods, yet longer dry periods between rain events,
leading to more drought.
At high latitudes and at high elevation, further changes
occur due to the loss of frozen water. Some of these are
resolved by the present generation of global climate
models (GCMs), and some changes can only be inferred
because they involve features such as glaciers, which
typically are not resolved or included in the models. The
warmer climate means that snow tends to start accu-
mulating later in the fall, and melt earlier in the spring.
Simulations assessed in this report show March to April
snow cover in the Northern Hemisphere is projected to
decrease by approximately 10 to 30% on average by
the end of this century, depending on the greenhouse
gas scenario. The earlier spring melt alters the timing
of peak springtime flow in rivers receiving snowmelt.
As a result, later flow rates will decrease, potentially
affecting water resource management. These features
appear in GCM simulations.
Loss of permafrost will allow moisture to seep more
deeply into the ground, but it will also allow the
ground to warm, which could enhance evapotranspiration. However, most current GCMs do not include all the pro-
cesses needed to simulate well permafrost changes. Studies analysing soils freezing or using GCM output to drive
more detailed land models suggest substantial permafrost loss by the end of this century. In addition, even though
current GCMs do not explicitly include glacier evolution, we can expect that glaciers will continue to recede, and
the volume of water they provide to rivers in the summer may dwindle in some locations as they disappear. Loss of
glaciers will also contribute to a reduction in springtime river flow. However, if annual mean precipitation increas-
es—either as snow or rain—then these results do not necessarily mean that annual mean river flow will decrease.
Land
evaporation
Wetter
Runoff
Drier
Wetter
Wetter
Drier
H
ad
l
e
y
FAQ 12.2, Figure 1 | Schematic diagram of projected changes in major com-
ponents of the water cycle. The blue arrows indicate major types of water move-
ment changes through the Earth’s climate system: poleward water transport by
extratropical winds, evaporation from the surface and runoff from the land to
the oceans. The shaded regions denote areas more likely to become drier or
wetter. Yellow arrows indicate an important atmospheric circulation change by
the Hadley Circulation, whose upward motion promotes tropical rainfall, while
suppressing subtropical rainfall. Model projections indicate that the Hadley
Circulation will shift its downward branch poleward in both the Northern and
Southern Hemispheres, with associated drying. Wetter conditions are projected
at high latitudes, because a warmer atmosphere will allow greater precipitation,
with greater movement of water into these regions.
1086
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
but there is strong agreement across the models over the direction of
change (Tebaldi et al., 2006; Goubanova and Li, 2007; Chen and Knut-
son, 2008; Haugen and Iversen, 2008; May, 2008b; Kysely and Berano-
va, 2009; Min et al., 2011; Sillmann et al., 2013). Regional details are
less robust in terms of the relative magnitude of changes but remain in
good accord across models in terms of the sign of the change and the
large-scale geographical patterns (Meehl et al., 2005a; CCSP, 2008a). In
semi-arid regions of the midlatitudes and subtropics such as the Medi-
terranean, the southwest USA, southwestern Australia, southern Africa
and a large portion of South America, the tendency manifested in the
majority of model simulations is for longer dry periods and is consist-
ent with the average decreases shown in Figure 12.22. Figure 12.26
shows projected percent changes in RX5day, the annual maximum of
consecutive 5-day precipitation over land regions obtained from the
CMIP5 models (Box 2.4, Table 1). Globally averaged end of 21st centu-
ry changes over land range from 5% (RCP2.6) to 20% (RCP8.5) more
precipitation during very wet 5-day periods. Results from the CMIP3
models are shown for comparison (see Section 12.4.9). Locally, the few
regions where this index of extreme precipitation decreases in the late
21st century RCP8.5 projection coincide with areas of robust decreases
in the mean precipitation of Figure 12.22.
Drought is discussed extensively in the SREX report (Seneviratne et al.,
2012) and the conclusions about future drought risk described there
based on CMIP3 models are reinforced by the CMIP5 models. As noted
in the SREX reports, assessments of changes in drought characteristics
with climate change should be made in the context of specific impacts
questions. The risk of future agricultural drought episodes is increased
in the regions of robust soil moisture decrease described in Section
12.4.5.3 and shown in Figure 12.23. Other measures in the literature of
future agricultural drought are largely focussed on the Palmer Drought
Severity Index (Wehner et al., 2011; Schwalm et al., 2012; Dai, 2013)
and project ‘extreme’ drought as the normal climatological state by
the end of the 21st century under the high emission scenarios in many
mid-latitude locations. However, this measure of agricultural drought
has been criticized as overly sensitive to increased temperatures due to
a simplified soil moisture model (Hoerling et al., 2012). The consecutive
dry-day index (CDD) is the length of the longest period of consecutive
days with precipitation less than 1 mm (Box 2.4, Table 1). CMIP5 pro-
jected changes in CDD over the 2081–2100 period under the RCP8.5
scenario (relative to the 1981–2000 reference period in common with
CMIP3) from the CMIP5 models are shown in Figure 12.26c and exhib-
it patterns similar to projected changes in both precipitation and soil
moisture (Sillmann et al., 2013). Substantial increases in this measure
of meteorological drought are projected in the Mediterranean, Central
America, Brazil, South Africa and Australia while decreases are project-
ed in high northern latitudes.
Truly rare precipitation events can cause very significant impacts. The
statistics of these events at the tails of the precipitation distribution
are well described by Extreme Value (EV) Theory although there are sig-
nificant biases in the direct comparison of gridded model output and
actual station data (Smith et al., 2009). There is also strong evidence
that model resolution plays a key role in replicating EV quantities esti-
mated from gridded observational data, suggesting that high-resolu-
tion models may provide somewhat more confidence in projection of
changes in rare precipitation events (Fowler et al., 2007a; Wehner et
al., 2011). Figure 12.27 shows the late 21st century changes per degree
Celsius in local warming in 20-year return values of annual maximum
daily precipitation relative to the late 20th century (left) and the asso-
ciated return periods of late 20th century 20-year return values at the
end of the 21st century from the CMIP5 models. Across future emission
scenarios, the global average of the CMIP5 multi-model median return
value sensitivity is an increase of 5.3% °C
–1
(Kharin et al., 2013). The
CMIP5 land average is close to the CMIP3 value of 4% °C
–1
report-
ed by Min et al. (2011) for a subset of CMIP3 models. Corresponding
with this change, the global average of return periods of late 20th
century 20-year return values is reduced from 20 years to 14 years for
a 1°C local warming. Return periods are projected to be reduced by
about 10 to 20% °C
–1
over the most of the mid-latitude land masses
with larger reductions over wet tropical regions (Kharin et al., 2013).
Hence, extreme precipitation events will very likely be more intense
Daily precipitation 20-yr RV change per 1°C warming
RP for present day 20-yr RV of daily precipitation
under 1°C warming
31 31
Figure 12.27 | (Left) The CMIP5 2081–2100 multi-model ensemble median percent change in 20-year return values of annual maximum daily precipitation per 1°C of local warm-
ing relative to the 1986–2005 reference period. (Right) The average 2081–2100 CMIP5 multi-model ensemble median of the return periods (years) of 1986–2005 20-year return
values of annual maximum daily precipitation corresponding to 1°C of local warming. Regions of no change would have return periods of 20 years.
1087
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
and more frequent in these regions in a warmer climate. Reductions in
return values (or equivalently, increases in return period) are confined
to convergent oceanic regions where circulation changes have reduced
the available water vapour.
Severe thunderstorms, associated with large hail, high winds, and tor-
nadoes, are another example of extreme weather associated with the
water cycle. The large-scale environments in which they occur are char-
acterized by large Convective Available Potential Energy (CAPE) and
deep tropospheric wind shear (Brooks et al., 2003; Brooks, 2009). Del
Genio et al. (2007), Trapp et al. (2007, 2009), and Van Klooster and Roe-
bber (2009) found a general increase in the energy and decrease in the
shear terms from the late 20th century to the late 21st century over the
USA using a variety of regional model simulations embedded in global
model SRES scenario simulations. The relative change between these
two competing factors would tend to favour more environments that
would support severe thunderstorms, providing storms are initiated.
Trapp et al. (2009), for example, found an increase in favourable thun-
derstorm conditions for all regions of the USA east of the Rocky Moun-
tains. Large variability in both the energy and shear terms means that
statistical significance is not reached until late in the 21st century under
high forcing scenarios. One way of assessing the possibility of a change
in the frequency of future thunderstorms is to look at historical records
of observed tornado, hail and wind occurrence with respect to the envi-
ronmental conditions (Brooks, 2013). This indicates that an increase in
the fraction of severe thunderstorms containing non-tornadic winds
would be consistent with the model projections of increased energy
and decreased shear, but there has not been enough research to make
a firm conclusion regarding future changes in frequency or magnitude.
Less work has been done on projected changes outside of the USA.
Marsh et al. (2009) found that mean energy decreased in the warm
season in Europe while it increased in the cool season. Even though the
energy decreases in the warm season, the number of days with favour-
able environments for severe thunderstorms increases because of an
increasing number of days with relatively large values of available
energy. For Europe, with the Mediterranean Sea and Sahara Desert to
the south, questions remain about changes in boundary layer moisture,
a main driver of the energy term. Niall and Walsh (2005) examined
changes in CAPE, which may be associated with hailstorm occurrence
in southeastern Australia using a global model, and found little change
under warmer conditions. Leslie et al. (2008) reconsidered the south-
eastern Australia hail question by nesting models with 1 km horizontal
grid spacing and using sophisticated microphysical parameterizations
and found an increase in the frequency of large hail by 2050 under the
SRES A1B scenario, but with extremely large internal variability in the
environments and hail size.
Overall, for all parts of the world studied, the results are suggestive of
a trend toward environments favouring more severe thunderstorms,
but the small number of analyses precludes any likelihood estimate of
this change.
12.4.6 Changes in Cryosphere
12.4.6.1 Changes in Sea Ice Cover
Based on the analysis of CMIP3 climate change simulations (e.g., Arzel
et al., 2006; Zhang and Walsh, 2006), the AR4 concludes that the Arctic
and Antarctic sea ice covers are projected to shrink in the 21st cen-
tury under all SRES scenarios, with a large range of model responses
(Meehl et al., 2007b). It also stresses that, in some projections, the
Arctic Ocean becomes almost entirely ice-free in late summer during
the second half of the 21st century. These conclusions were confirmed
by further analyses of the CMIP3 archives (e.g., Stroeve et al., 2007;
Bracegirdle et al., 2008; Lefebvre and Goosse, 2008; Boé et al., 2009b;
Sen Gupta et al., 2009; Wang and Overland, 2009; Zhang, 2010b; NRC,
2011; Körper et al., 2013). Figures 12.28 and 12.29 and the studies of
Maksym et al. (2012), Massonnet et al. (2012), Stroeve et al. (2012) and
Wang and Overland (2012) show that the CMIP5 AOGCMs/ESMs as a
group also project decreases in sea ice extent through the end of this
century in both hemispheres under all RCPs. However, as in the case of
CMIP3, the inter-model spread is considerable.
In the NH, in accordance with CMIP3 results, the absolute rate of
decrease of the CMIP5 multi-model mean sea ice areal coverage is
greatest in September. The reduction in sea ice extent between the
time periods 1986–2005 and 2081–2100 for the CMIP5 multi-model
average ranges from 8% for RCP2.6 to 34% for RCP8.5 in February
and from 43% for RCP2.6 to 94% for RCP8.5 in September. Medium
confidence is attached to these values as projections of sea ice extent
decline in the real world due to errors in the simulation of present-day
sea ice extent (mean and trends—see Section 9.4.3) and because
of the large spread of model responses. About 90% of the available
CMIP5 models reach nearly ice-free conditions (sea ice extent less than
1 × 10
6
km
2
for at least 5 consecutive years) during September in the
Arctic before 2100 under RCP8.5 (about 45% under RCP4.5). By the
end of the 21st century, the decrease in multi-model mean sea ice
volume ranges from 29% for RCP2.6 to 73% for RCP8.5 in February
and from 54% for RCP2.6 to 96% for RCP8.5 in September. Medium
confidence is attached to these values as projections of the real world
sea ice volume. In February, these percentages are much higher than
the corresponding ones for sea ice extent, which is indicative of a sub-
stantial sea ice thinning.
A frequent criticism of the CMIP3 models is that, as a group, they
strongly underestimate the rapid decline in summer Arctic sea ice
extent observed during the past few decades (e.g., Stroeve et al., 2007;
Winton, 2011), which suggests that the CMIP3 projections of summer
Arctic sea ice areal coverage might be too conservative. As shown in
Section 9.4.3 and Figure 12.28b, the magnitude of the CMIP5 mul-
ti-model mean trend in September Arctic sea ice extent over the satel-
lite era is more consistent with, but still underestimates, the observed
one (see also Massonnet et al., 2012; Stroeve et al., 2012; Wang and
Overland, 2012; Overland and Wang, 2013). Owing to the shortness of
the observational record, it is difficult to ascertain the relative influ-
ence of natural variability on this trend. This hinders the comparison
between modelled and observed trends, and hence the estimate of the
sensitivity of the September Arctic sea ice extent to global surface tem-
perature change (i.e., the decrease in sea ice extent per degree global
1088
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
warming) (Kay et al., 2011; Winton, 2011; Mahlstein and Knutti, 2012).
This sensitivity may be crucial for determining future sea ice losses.
Indeed, a clear relationship exists at longer than decadal time scales
in climate change simulations between the annual mean or September
mean Arctic sea ice extent and the annual mean global surface tem-
perature change for ice extents larger than ~1 × 10
6
km
2
(e.g., Ridley
et al., 2007; Zhang, 2010b; NRC, 2011; Winton, 2011; Mahlstein and
Knutti, 2012). This relationship is illustrated in Figure 12.30 for both
CMIP3 and CMIP5 models. From this figure, it can be seen that the sea
ice sensitivity varies significantly from model to model and is generally
larger and in better agreement among models in CMIP5.
A complete and detailed explanation for what controls the range of
Arctic sea ice responses in models over the 21st century remains elu-
sive, but the Arctic sea ice provides an example where process-based
constraints can be used to reduce the spread of model projections
Year
Sea ice extent change (10
6
km
2
)
Northern Hemisphere February
Satellite obs. 1986−2005 avg: 15.5 x 10
6
km
2
CMIP5 historical 1986−2005 avg: 15.9 x10
6
km
2
Historical (39)
Observations
a)
RCP2.6 (29)
RCP4.5 (39)
RCP6.0 (21)
RCP8.5 (37)
1960 1980 2000 2020 2040 2060 2080 2100
−7
−6
−5
−4
−3
−2
−1
0
1
2
3
Year
Sea ice extent change (10
6
km
2
)
Northern Hemisphere September
Satellite obs. 1986−2005 avg: 7.1 x10
6
km
2
CMIP5 historical 1986−2005 avg: 6.6 x10
6
km
2
Historical (39)
Observations
b)
RCP2.6 (29)
b)
RCP4.5 (39)
RCP6.0 (21)
RCP8.5 (37)
1960 1980 2000 2020 2040 2060 2080 2100
−7
−6
−5
−4
−3
−2
−1
0
1
2
3
Year
Sea ice extent change (10
6
km
2
)
Southern Hemisphere February
Satellite obs. 1986−2005 avg: 3.3 x10
6
km
2
CMIP5 historical 1986−2005 avg: 3.0 x10
6
km
2
Historical (39)
Observations
c)
RCP2.6 (29)
RCP4.5 (39)
RCP6.0 (21)
RCP8.5 (37)
1960 1980 2000 2020 2040 2060 2080 2100
−7
−6
−5
−4
−3
−2
−1
0
1
2
3
Year
Sea ice extent change (10
6
km
2
)
Southern Hemisphere September
Satellite obs. 1986−2005 avg: 19.0 x10
6
km
2
CMIP5 historical 1986−2005 avg: 17.8 x10
6
km
2
Historical (39)
Observations
d)
RCP2.6 (29)
RCP4.5 (39)
RCP6.0 (21)
RCP8.5 (37)
1960 1980 2000 2020 2040 2060 2080 2100
−7
−6
−5
−4
−3
−2
−1
0
1
2
3
Figure 12.28 | Changes in sea ice extent as simulated by CMIP5 models over the second half of the 20th century and the whole 21st century under RCP2.6, RCP4.5, RCP6.0 and
RCP8.5 for (a) Northern Hemisphere February, (b) Northern Hemisphere September, (c) Southern Hemisphere February and (d) Southern Hemisphere September. The solid curves
show the multi-model means and the shading denotes the 5 to 95% range of the ensemble. The vertical line marks the end of CMIP5 historical climate change simulations. One
ensemble member per model is taken into account in the analysis. Sea ice extent is defined as the total ocean area where sea ice concentration exceeds 15% and is calculated on
the original model grids. Changes are relative to the reference period 1986–2005. The number of models available for each RCP is given in the legend. Also plotted (solid green
curves) are the satellite data of Comiso and Nishio (2008, updated 2012) over 1979–2012.
(Overland et al., 2011; Collins et al., 2012; Hodson et al., 2012). For
CMIP3 models, results indicate that the changes in Arctic sea ice mass
budget over the 21st century are related to the late 20th century mean
sea ice thickness distribution (Holland et al., 2010), average sea ice
thickness (Bitz, 2008; Hodson et al., 2012), fraction of thin ice cover
(Boé et al., 2009b) and oceanic heat transport to the Arctic (Mahlstein
and Knutti, 2011). For CMIP5 models, Massonnet et al. (2012) showed
that the time needed for the September Arctic sea ice areal coverage to
drop below a certain threshold is highly correlated with the September
sea ice extent and annual mean sea ice volume averaged over the past
several decades (Figure 12.31a, b). The timing of a seasonally ice-free
Arctic Ocean or the fraction of remaining sea ice in September at any
time during the 21st century were also found to correlate with the
past trend in September Arctic sea ice extent and the amplitude of the
mean seasonal cycle of sea ice extent (Boé et al., 2009b; Collins et al.,
2012; Massonnet et al., 2012) (Figure 12.31c, d). All these empirical
1089
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Figure 12.29 | February and September CMIP5 multi-model mean sea ice concentrations (%) in the Northern and Southern Hemispheres for the periods (a) 1986–2005, (b)
2081–2100 under RCP4.5 and (c) 2081–2100 under RCP8.5. The model sea ice concentrations are interpolated onto a 1° × 1° regular grid. One ensemble member per model is
taken into account in the analysis, and the multi-model mean sea ice concentration is shown where it is larger than 15%. The number of models available for each RCP is given in
parentheses. The pink lines indicate the observed 15% sea ice concentration limits averaged over 1986–2005 (Comiso and Nishio, 2008, updated 2012).
relationships can be understood on simple physical grounds (see the
aforementioned references for details).
These results lend support for weighting/recalibrating the models
based on their present-day Arctic sea ice simulations. Today, the opti-
mal approach for constraining sea ice projections from climate models
is unclear, although one notes that these methods should have a
credible underlying physical basis in order to increase confidence in
their results (see Section 12.2). In addition, they should account for
February September September February
February September September February
February September September February
a) 1986−2005 average (39)
b) 2081−2100 average, RCP4.5 (39)
c) 2081−2100 average, RCP8.5 (37)
(%)
0 20 40 60 80 100
the potentially large imprint of natural variability on both observations
and model simulations when these two sources of information are to
be compared (see Section 9.8.3). This latter point is particularly critical
if the past sea ice trend or sensitivity is used in performance metrics
given the relatively short observational period (Kay et al., 2011; Over-
land et al., 2011; Mahlstein and Knutti, 2012; Massonnet et al., 2012;
Stroeve et al., 2012). A number of studies have applied such metrics
to the CMIP3 and CMIP5 models. Stroeve et al. (2007) and Stroeve et
al. (2012) rejected several CMIP3 and CMIP5 models, respectively, on
1090
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
the basis of their simulated late 20th century mean September Arctic
sea ice extent. Wang and Overland (2009) selected a subset of CMIP3
models (and Wang and Overland (2012) did the same for the CMIP5
models) based on their fidelity to the observed mean seasonal cycle of
Arctic sea ice extent in the late 20th century and then scaled the chosen
models to the recently observed September sea ice extent. Zhang
(2010b) retained a number of CMIP3 models based on the regression
between summer sea ice loss and Arctic surface temperature change.
Boé et al. (2009b) and Mahlstein and Knutti (2012) did not perform a
model selection but rather recalibrated the CMIP3 Arctic sea ice projec-
tions on available observations of September Arctic sea ice trend and
sensitivity to global surface temperature change, respectively. Finally,
Massonnet et al. (2012) selected a subset of CMIP5 models on the
basis of the four relationships illustrated in Figure 12.31a–d.
These various methods all suggest a faster rate of summer Arctic sea
ice decline than the multi-model mean. Although they individually
provide a reduced range for the year of near disappearance of the
September Arctic sea ice compared to the original CMIP3/CMIP5 mul-
ti-model ensemble, they lead to different timings (Overland and Wang,
2013). Consequently, the time interval obtained when combining all
these studies remains wide: 2020–2100
+
(2100
+
= not before 2100)
for the SRES A1B scenario and RCP4.5 (Stroeve et al., 2007, 2012; Boé
et al., 2009b; Wang and Overland, 2009, 2012; Zhang, 2010b; Masson-
net et al., 2012) and 2020–2060 for RCP8.5 (Massonnet et al., 2012;
Wang and Overland, 2012). The method proposed by Massonnet et
al. (2012) is applied here to the full set of models that provided the
CMIP5 database with sea ice output. The natural variability of each
of the four diagnostics shown in Figure 12.31a–d is first estimated
by averaging over all available models with more than one ensemble
member the diagnostic standard deviations derived from the model
CMIP5 (b)
Annual mean global surface warming (°C)
September Arctic sea ice extent (10
6
km
2
)
0 1 2 3 4 5
0
2
4
6
8
10
12
CMIP3 (a)
Annual mean global surface warming (°C)
September Arctic sea ice extent (10
6
km
2
)
0 1 2 3 4 5
0
2
4
6
8
10
12
Figure 12.30 | September Arctic sea ice extent as a function of the annual mean global surface warming relative to the period 1986–2005 for (a) CMIP3 models (all SRES sce-
narios) and (b) CMIP5 models (all RCPs). The ice extents and global temperatures are computed on a common latitude-longitude grid for CMIP3 and on the original model grids for
CMIP5. One ensemble member per model is taken into account in the analysis. A 21-year running mean is applied to the model output. The full black circle and vertical bar on the
left-hand side of the y-axis indicate the mean and ±2 standard deviations about the mean of the observed September Arctic sea ice extent over 1986–2005, respectively (Comiso
and Nishio, 2008, updated 2012). The horizontal line corresponds to a nearly ice-free Arctic Ocean in September.
ensemble members. Then, for each model, a ±2 standard deviation
interval is constructed around the ensemble mean or single realization
of the diagnostic considered. A model is retained if, for each diagnostic,
either this interval overlaps a ±20% interval around the observed/rea-
nalysed value of the diagnostic or at least one ensemble member from
that model gives a value for the diagnostic that falls within ±20% of
the observational/reanalysed data. The outcome is displayed in Figure
12.31e for RCP8.5. Among the five selected models (ACCESS1.0,
ACCESS1.3, GFDL-CM3, IPSL-CM5A-MR, MPI-ESM-MR), four project
a nearly ice-free Arctic Ocean in September before 2050 (2080) for
RCP8.5 (RCP4.5), the earliest and latest years of near disappearance
of the sea ice pack being about 2040 and about 2060 (about 2040
and 2100
+
), respectively. It should be mentioned that Maslowski et al.
(2012) projected that it would take only until about 2016 to reach a
nearly ice-free Arctic Ocean in summer, based on a linear extrapolation
into the future of the recent sea ice volume trend from a hindcast sim-
ulation conducted with a regional model of the Arctic sea ice–ocean
system. However, such an extrapolation approach is problematic as it
ignores the negative feedbacks that can occur when the sea ice cover
becomes thin (e.g., Bitz and Roe, 2004; Notz, 2009) and neglects the
effect of year-to-year or longer-term variability (Overland and Wang,
2013). Mahlstein and Knutti (2012) encompassed the dependence of
sea ice projections on the forcing scenario by determining the annual
mean global surface warming threshold for nearly ice-free conditions
in September. Their best estimate of ~2°C above the present derived
from both CMIP3 models and observations is consistent with the 1.6
to 2.1°C range (mean value: 1.9°C) obtained from the CMIP5 model
subset shown in Figure 12.31e (see also Figure 12.30b). The reduction
in September Arctic sea ice extent by the end of the 21st century, aver-
aged over this subset of models, ranges from 56% for RCP2.6 to 100%
for RCP8.5.
1091
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Figure 12.31 | (a–d) First year during which the September Arctic sea ice extent falls below 1 × 10
6
km
2
in CMIP5 climate projections (37 models, RCP8.5) as a function of (a)
the September Arctic sea ice extent averaged over 1986–2005, (b) the annual mean Arctic sea ice volume averaged over 1986–2005, (c) the amplitude of the 1986–2005 mean
seasonal cycle of Arctic sea ice extent and (d) the trend in September Arctic sea ice extent over 1979–2012. The sea ice diagnostics displayed are calculated on the original model
grids. The correlations and one-tailed p-values are computed from the multi-member means for models with several ensemble members (coloured crosses), but the ensemble mem-
bers of individual models are also depicted (coloured dots). The vertical solid and dashed lines show the corresponding observations or bias-adjusted PIOMAS (Pan-Arctic Ice-Ocean
Modelling and Assimilation System) reanalysis data (a, c and d: Comiso and Nishio, 2008, updated 2012; b: Schweiger et al., 2011) and the ±20% interval around these data,
respectively. (e) Time series of September Arctic sea ice extent (5-year running mean) as simulated by all CMIP5 models and their ensemble members under RCP8.5 (thin curves).
The thick, coloured curves correspond to a subset of five CMIP5 models selected on the basis of panels a–d following Massonnet et al. (2012) (see text for details). Note that each
of these models provides only one ensemble member for RCP8.5.
RCP8.5, correlation = 0.82, p = 1e−09
September Arctic sea ice extent
averaged over 1986−2005 (10
6
km
2
)
a)
2 3 4 5 6 7 8 9 10 11
2000
2020
2040
2060
2080
2100
RCP8.5, correlation = 0.64, p = 2e−05
Annual mean Arctic sea ice volume
averaged over 1986−2005 (10
3
km
3
)
b)
10 15 20 25 30 35 40 45
2000
2020
2040
2060
2080
2100
RCP8.5, correlation = −0.53, p = 0.0007
Amplitude of the seasonal cycle of Arctic sea ice extent
averaged over 1986−2005 (10
6
km
2
)
First year of near disappearance of September Arctic sea ice
c)
0 2 4 6 8 10 12 14 16 18
2000
2020
2040
2060
2080
2100
RCP8.5, correlation = 0.48, p = 0.002
Trend in September Arctic sea ice extent
over 1979−2012 (10
3
km
2
/decade)
d)
−1600 −1200 −800 −400 0
2000
2020
2040
2060
2080
2100
Year
September Arctic sea ice extent (10
6
km
2
)
RCP8.5
e)
2010 2020 2030 2040 2050 2060 2070 2080 2090 2100
0
1
2
3
4
5
6
7
8
9
10
11
12
1092
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
In light of all these results, it is very likely that the Arctic sea ice cover
will continue to shrink and thin all year round during the 21st century
as the annual mean global surface temperature rises. It is also likely
that the Arctic Ocean will become nearly ice-free in September before
the middle of the century for high GHG emissions such as those corre-
sponding to RCP8.5 (medium confidence). The potential irreversibility
of the Arctic sea ice loss and the possibility of an abrupt transition
toward an ice-free Arctic Ocean are discussed in Section 12.5.5.7.
In the SH, the decrease in sea ice extent between 1986–2005 and
2081–2100 projected by the CMIP5 models as a group varies from
16% for RCP2.6 to 67% for RCP8.5 in February and from 8% to 30%
in September. In contrast with the NH, the absolute rate of decline is
greatest in wintertime. Eisenman et al. (2011) argue that this hemi-
spheric asymmetry in the seasonality of sea ice loss is fundamentally
related to the geometry of coastlines. For each forcing scenario, the
relative changes in multi-model mean February and September Antarc-
tic sea ice volumes by the end of the century are of the same order as
the corresponding ones for sea ice extent. About 75% of the available
CMIP5 models reach a nearly ice-free state in February within this cen-
tury under RCP8.5 (about 60% under RCP4.5). For RCP8.5, only small
portions of the Weddell and Ross Seas stay ice-covered in February
during 2081–2100 in those models that do not project a seasonally
ice-free Southern Ocean (see Figure 12.29c). Nonetheless, there is low
confidence in these Antarctic sea ice projections because of the wide
range of model responses and the inability of almost all of the models
to reproduce the mean seasonal cycle, interannual variability and over-
all increase of the Antarctic sea ice areal coverage observed during the
satellite era (see Section 9.4.3; Maksym et al., 2012; Turner et al., 2013;
Zunz et al., 2013).
12.4.6.2 Changes in Snow Cover and Frozen Ground
Excluding ice sheets and glaciers, analyses of seasonal snow cover
changes generally focus on the NH, where the configuration of the
continents on the Earth induces a larger maximum seasonal snow
cover extent (SCE) and a larger sensitivity of SCE to climate changes.
Seasonal snow cover extent and snow water equivalent (SWE) respond
to both temperature and precipitation. At the beginning and the end
of the snow season, SCE decreases are closely linked to a shortening
of the seasonal snow cover duration, while SWE is more sensitive to
snowfall amount (Brown and Mote, 2009). Future widespread reduc-
tions of SCE, particularly in spring, are simulated by the CMIP3 models
(Roesch, 2006; Brown and Mote, 2009) and confirmed by the CMIP5
ensemble (Brutel-Vuilmet et al., 2013). The NH spring (March-April
average) snow cover area changes are coherent in the CMIP5 models
although there is considerable scatter. Relative to the 1986–2005 ref-
erence period, the CMIP5 models simulate a weak decrease of about
7 ± 4% (one-σ inter-model dispersion) for RCP2.6 during the last two
decades of the 21st century, while SCE decreases of about 13 ± 4% are
simulated for RCP4.5, 15 ± 5% for RCP6.0, and 25 ± 8% for RCP8.5
(Figure 12.32). There is medium confidence in these numbers because
of the considerable inter-model scatter mentioned above and because
snow processes in global climate models are strongly simplified.
Projections for the change in annual maximum SWE are more mixed.
Warming decreases SWE both by reducing the fraction of precipitation
that falls as snow and by increasing snowmelt, but projected increas-
es in precipitation over much of the northern high latitudes during
winter months act to increase snow amounts. Whether snow cover-
ing the ground will become thicker or thinner depends on the balance
between these competing factors. Both in the CMIP3 (Räisänen, 2008)
and in the CMIP5 models (Brutel-Vuilmet et al., 2013), annual maxi-
mum SWE tends to increase or only marginally decrease in the coldest
20
0
-20
-40
(%)
Snow cover extent change
Figure 12.32 | Northern Hemisphere spring (March to April average) snow cover
extent change (in %) in the CMIP5 ensemble, relative to the simulated extent for the
1986–2005 reference period. Thick lines mark the multi-model average, shading indi-
cates the inter-model spread (one standard deviation). The observed March to April
average snow cover extent for the 1986–2005 reference period is 32.6·10
6
km
2
(Brown
and Robinson, 2011).
Near-surface permafrost area
(10
6
km
2
)
Figure 12.33 | Northern Hemisphere near-surface permafrost area, diagnosed for
the available CMIP5 models by Slater and Lawrence (2013) following Nelson and Out-
calt (1987) and using 20-year average bias-corrected monthly surface air temperatures
and snow depths. Thick lines: multi-model average. Shading and thin lines indicate the
inter-model spread (one standard deviation). The black line for the historical period is
diagnosed from the average of the European Centre for Medium range Weather Fore-
cast (ECMWF) reanalysis of the global atmosphere and surface conditions (ERA), Japa-
nese ReAnalysis (JRA), Modern Era Retrospective-analysis for Research and Applications
(MERRA) and Climate Forecast System Reanalysis and Reforecast (CFSRR) reanalyses
(Slater and Lawrence, 2013). Estimated present permafrost extent is between 12 and
17 million km
2
(Zhang et al., 2000).
1093
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
regions, while annual maximum SWE decreases are strong closer to the
southern limit of the seasonally snow-covered area.
It is thus very likely (high confidence) that by the end of the 21st centu-
ry, NH spring snow cover extent will be substantially lower than today
if anthropogenic climate forcing is similar to the stronger scenarios
considered here. Conversely, there is only medium confidence in the
latitudinal pattern of annual maximum SWE changes (increase or little
change in the coldest regions, stronger decrease further to the South)
because annual maximum SWE is influenced by competing factors
(earlier melt onset, higher solid precipitation rates in some regions).
The strong projected warming across the northern high latitudes in
climate model simulations has implications for frozen ground. Recent
projections of the extent of near-surface permafrost (see Glossary)
degradation continue to vary widely depending on the underlying
climate forcing scenario and model physics, but virtually all of them
indicate substantial near-surface permafrost degradation and thaw
depth deepening over much of the permafrost area (Saito et al., 2007;
Lawrence et al., 2008a, 2012; Koven et al., 2011, 2013; Eliseev et al.,
2013; Slater and Lawrence, 2013). Permafrost at greater depths is less
directly relevant to the surface energy and water balance, and its deg-
radation naturally occurs much more slowly (Delisle, 2007). Climate
models are beginning to represent permafrost physical processes and
properties more accurately (Alexeev et al., 2007; Nicolsky et al., 2007;
Lawrence et al., 2008a; Rinke et al., 2008; Koven et al., 2009; Gout-
tevin et al., 2012), but there are large disagreements in the calculation
of current frozen soil extent and active layer depth due to differenc-
es in the land model physics in the CMIP5 ensemble (Koven et al.,
2013). The projected changes in permafrost are a response not only
to warming, but also to changes in snow conditions because snow
properties and their seasonal evolution exert significant control on soil
thermal state (Zhang, 2005; Lawrence and Slater, 2010; Shkolnik et
al., 2010; Koven et al., 2013). Applying the surface frost index method
(Nelson and Outcalt, 1987) to coupled climate model anomalies from
the CMIP5 models (Slater and Lawrence, 2013) yields a reduction of
the diagnosed 2080–2099 near-surface permafrost area (continuous
plus discontinuous near-surface permafrost) by 37 ± 11% (RCP2.6),
51 ± 13% (RCP4.5), 58 ± 13% (RCP6.0), and 81±12% (RCP8.5), com-
pared to the 1986–2005 diagnosed near-surface permafrost area, with
medium confidence in the numbers as such because of the strongly
simplified soil physical processes in current-generation global climate
models (Figure 12.33). The uncertainty range given here is the 1-σ
inter-model dispersion. Applying directly the model output to diag-
nose permafrost extent and its changes over the 21st century yields
similar relative changes (Koven et al., 2013). In summary, based on
high agreement across CMIP5 and older model projections, fundamen-
tal process understanding, and paleoclimatic evidence (e.g., Vaks et
al., 2013), it appears virtually certain (high confidence) that near-sur-
face permafrost extent will shrink as global climate warms. However,
the amplitude of the projected reductions of near-surface permafrost
extent not only depends on the emission scenario and the global cli-
mate model response, but also very much on the permafrost-related
soil processes taken into account in the models.
12.4.7 Changes in the Ocean
12.4.7.1 Sea Surface Temperature, Salinity and Ocean
Heat Content
Projected increase of SST and heat content over the next two decades
is relatively insensitive to the emissions trajectory. However, projected
outcomes diverge as the 21st century progresses. When SSTs increase
as a result of external forcing, the interior water masses respond to
the integrated signal at the surface, which is then propagated down to
greater depth (Gleckler et al., 2006; Gregory, 2010). Changes in glob-
ally averaged ocean heat content currently account for about 90% of
the change in global energy inventory since 1970 (see Box 3.1). Heat is
transported within the interior of the ocean by its large-scale general
circulation and by smaller-scale mixing processes. Changes in trans-
ports lead to redistribution of existing heat content and can cause local
cooling even though the global mean heat content is rising (Banks and
Gregory, 2006; Lowe and Gregory, 2006; Xie and Vallis, 2012).
Figure 12.12 shows the multi-model mean projections of zonally aver-
aged ocean temperature change under three emission scenarios. The
differences in projected ocean temperature changes for different RCPs
manifest themselves more markedly as the century progresses. The
largest warming is found in the top few hundred metres of the subtrop-
ical gyres, similar to the observed pattern of ocean temperature chang-
es (Levitus et al., 2012, see also Section 3.2). Surface warming varies
considerably between the emission scenarios ranging from about 1°C
(RCP2.6) to more than 3°C in RCP8.5. Mixing and advection processes
gradually transfer the additional heat to deeper levels of about 2000
m at the end of the 21st century. Depending on the emission scenario,
global ocean warming between 0.5°C (RCP2.6) and 1.5°C (RCP8.5)
will reach a depth of about 1 km by the end of the century. The stron-
gest warming signal is found at the surface in subtropical and tropical
regions. At depth the warming is most pronounced in the Southern
Ocean. From an energy point of view, for RCP4.5 by the end of the 21st
century, half of the energy taken up by the ocean is in the uppermost
700 m, and 85% is in the uppermost 2000 m.
In addition to the upper-level warming, the patterns are further char-
acterized by a slight cooling in parts of the northern mid- and high
latitudes below 1000 m and a pronounced heat uptake in the deep
Southern Ocean at the end of the 21st century. The cooling may be
linked to the projected decrease of the strength of the AMOC (see Sec-
tion 12.4.7.2; 13.4.1; Banks and Gregory, 2006).
The response of ocean temperatures to external forcing comprises
mainly two time scales: a relatively fast adjustment of the ocean mixed
layer and the slow response of the deep ocean (Hansen et al., 1985;
Knutti et al., 2008a; Held et al., 2010). Simulations with coupled ocean–
atmosphere GCMs suggest time-scales of several millennia until the
deep ocean is in equilibrium with the external forcing (Stouffer, 2004;
Hansen et al., 2011; Li et al., 2013a). Thus, the long time-scale of the
ocean response to external forcing implies an additional commitment
to warming for many centuries when GHG emissions are decreased or
concentrations kept constant (see Section 12.5.2). Further assessment
of ocean heat uptake and its relationship to projections of sea level rise
is presented in Section 13.4.1.
1094
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Durack and Wijffels (2010) and Durack et al. (2012) examined trends
in global sea surface salinity (SSS) changes over the period 1950–
2008. Their analysis revealed strong, spatially coherent trends in SSS
over much of the global ocean, with a pattern that bears striking
resemblance to the climatological SSS field and is associated with an
intensification of the global water cycle (see Sections 3.3.2.1, 10.4.2
and 12.4.5). The CMIP5 climate model projections available suggest
that high SSS subtropical regions that are dominated by net evapora-
tion are typically getting more saline; lower SSS regions at high lati-
tudes are typically getting fresher. They also suggest a continuation
of this trend in the Atlantic where subtropical surface waters become
more saline as the century progresses (Figure 12.34) (see also Terray et
al., 2012). At the same time, the North Pacific is projected to become
less saline.
12.4.7.2 Atlantic Meridional Overturning
Almost all climate model projections reveal an increase of high latitude
temperature and high latitude precipitation (Meehl et al., 2007b). Both
of these effects tend to make the high latitude surface waters lighter
and hence increase their stability. As seen in Figure 12.35, all models
show a weakening of the AMOC over the course of the 21st century
(see Section 12.5.5.2 for further analysis). Projected changes in the
strength of the AMOC at high latitudes appear stronger in Geophysical
Fluid Dynamics Laboratory (GFDL) CM2.1 when density is used as a
vertical coordinate instead of depth (Zhang, 2010a). Once the RF is sta-
bilized, the AMOC recovers, but in some models to less than its pre-in-
dustrial level. The recovery may include a significant overshoot (i.e., a
weaker circulation may persist) if the anthropogenic RF is eliminated
(Wu et al., 2011a). Gregory et al. (2005) found that for all eleven models
Figure 12.34 | Projected sea surface salinity differences 2081–2100 for RCP8.5 rela-
tive to 1986–2005 from CMIP5 models. Hatching indicates regions where the multi-
model mean change is less than one standard deviation of internal variability. Stippling
indicates regions where the multi-model mean change is greater than two standard
deviations of internal variability and where at least 90% of the models agree on the
sign of change (see Box 12.1). The number of CMIP5 models used is indicated in the
upper right corner.
Atlantic Meridional Overturning Circulation at 30
o
N
(Sv)(Sv)
Figure 12.35 | Multi-model projections of Atlantic Meridional Overturning Circulation (AMOC) strength at 30°N from 1850 through to the end of the RCP extensions. Results are
based on a small number of CMIP5 models available. Curves show results from only the first member of the submitted ensemble of experiments.
1095
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
analysed (six from CMIP2/3 and five EMICs), the AMOC reduction was
caused more by changes in surface heat flux than changes in surface
freshwater flux. They further found that models with a stronger AMOC
in their control run exhibited a larger weakening (see also Gregory and
Tailleux, 2011).
Based on the assessment of the CMIP5 RCP simulations and on our
understanding gleaned from analysis of CMIP3 models, observations
and our understanding of physical mechanisms, it is very likely that the
AMOC will weaken over the 21st century. Best estimates and ranges
for the reduction from CMIP5 are 11% (1 to 24%) in RCP2.6 and 34%
(12 to 54%) in RCP8.5. There is low confidence in assessing the evolu-
tion of the AMOC beyond the 21st century.
12.4.7.3 Southern Ocean
A dominant and robust feature of the CMIP3 climate projections
assessed in AR4 is the weaker surface warming at the end of the 21st
century in the Southern Ocean area compared to the global mean. Fur-
thermore, the Antarctic Circumpolar Current (ACC) moves southward
in most of the climate projections analysed in response to the simulat-
ed southward shift and strengthening of the SH mid-latitude westerlies
(Meehl et al., 2007b).
The additional analyses of the CMIP3 model output performed since
the release of AR4 confirm and refine the earlier findings. The displace-
ment and intensification of the mid-latitude westerlies contribute to a
large warming between 40°S and 60°S from the surface to mid-depths
(Fyfe et al., 2007; Sen Gupta et al., 2009). Part of this warming has
been attributed to the southward translation of the Southern Ocean
current system (Sen Gupta et al., 2009). Moreover, the wind changes
influence the surface temperature through modifications of the latent
and sensible heat fluxes and force a larger northward Ekman trans-
port of relatively cold polar surface water (Screen et al., 2010). This
also leads to a stronger upwelling that brings southward and upward
relatively warm and salty deep water, resulting in a subsurface salinity
increase at mid-depths south of 50°S (Sen Gupta et al., 2009; Screen
et al., 2010).
Overall, CMIP3 climate projections exhibit a decrease in mixed layer
depth at southern mid- and high latitudes by the end of the 21st centu-
ry. This feature is a consequence of the enhanced stratification resulting
from surface warming and freshening (Lefebvre and Goosse, 2008; Sen
Gupta et al., 2009; Capotondi et al., 2012). Despite large inter-mod-
el differences, there is a robust weakening of Antarctic Bottom Water
production and its northward outflow, which is consistent with the
decrease in surface density and is manifest as a warming signal close
to the Antarctic margin that reaches abyssal depths (Sen Gupta et al.,
2009).
In the vicinity of the Antarctic ice sheet, CMIP3 models project an aver-
age warming of ~0.5C° at depths of 200–500 m in 2091–2100 com-
pared to 1991–2000 for the SRES A1B scenario, which has the poten-
tial to impact the mass balance of ice shelves (Yin et al., 2011). More
detailed regional modelling using the SRES A1B scenario indicates
that a redirection of the coastal current into the cavities underlying
the Filchner-Ronne ice shelf during the second half of the 21st century
might enhance the average basal melting rate there from 0.2 m yr
–1
to
almost 4 m yr
–1
(Hellmer et al., 2012; see Section 13.4.4.2).
There are very few published analyses of CMIP5 climate projections
focusing on the Southern Ocean. Meijers et al. (2012) found a wide
variety of ACC responses to climate warming scenarios across CMIP5
models. Models show a high correlation between the changes in
ACC strength and position, with a southward (northward) shift of the
ACC core as the ACC gets stronger (weaker). No clear relationship
between future changes in wind stress and ACC strength was identi-
fied, while the weakening of the ACC transport simulated at the end
of the 21st century by many models was found to correlate with the
strong decrease in the surface heat and freshwater fluxes in the ACC
region (Meijers et al., 2012; Downes and Hogg, 2013). In agreement
with the CMIP3 assessment (Sen Gupta et al., 2009), subtropical gyres
generally strengthen under RCP4.5 and RCP8.5 and all expand south-
ward, inducing a southward shift of the northern boundary of the ACC
at most longitudes in the majority of CMIP5 models (Meijers et al.,
2012). As in CMIP3 climate projections, an overall shallowing of the
deep mixed layers that develop on the northern edge of the ACC in
winter is observed, with larger shallowing simulated by models with
deeper mixed layers during 1976–2005 (Sallée et al., 2013a). Sallée
et al. (2013b) reported a warming of all mode, intermediate and deep
water masses in the Southern Ocean. The largest temperature increase
is found in mode and intermediate water layers. Consistently with
CMIP3 projections (Downes et al., 2010), these water layers experience
a freshening, whereas bottom water becomes slightly saltier. Finally,
Sallée et al. (2013b) noted an enhanced upwelling of circumpolar deep
water and an increased subduction of intermediate water that are
nearly balanced by interior processes (diapycnal fluxes).
A number of studies suggest that oceanic mesoscale eddies might
influence the response of the Southern Ocean circulation, meridional
heat transport and deep water formation to changes in wind stress and
surface buoyancy flux (Böning et al., 2008; Farneti et al., 2010; Downes
et al., 2011; Farneti and Gent, 2011; Saenko et al., 2012; Spence et al.,
2012). These eddies are not explicitly resolved in climate models and
their role in future circulation changes still needs to be precisely quan-
tified. Some of the CMIP5 models have output the meridional overturn-
ing due to the Eulerian mean circulation and that induced by parame-
terized eddies, thus providing a quantitative estimate of the role of the
mesoscale circulation in a warming climate. On this basis, Downes and
Hogg (2013) found that, under RCP8.5, the strengthening (weakening)
of the upper (lower) Eulerian mean meridional overturning cell in the
Southern Ocean is significantly correlated with the increased overlying
wind stress and surface warming and is partly compensated at best by
changes in eddy-induced overturning.
None of the CMIP3 and CMIP5 models include an interactive ice sheet
component. When climate-ice sheet interactions are accounted for in
an EMIC under a 4 × CO
2
scenario, the meltwater flux from the Antarc-
tic ice sheet further reduces the surface density close to Antarctica and
the rate of Antarctic Bottom Water formation. This ultimately results
in a smaller surface warming at high southern latitudes compared to
a simulation in which the freshwater flux from the melting ice sheet is
not taken into account (Swingedouw et al., 2008). Nevertheless, in this
study, this effect becomes significant only after more than one century.
1096
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
12.4.8 Changes Associated with Carbon Cycle Feedbacks
and Vegetation Cover
Climate change may affect the global biogeochemical cycles changing
the magnitude of the natural sources and sinks of major GHGs. Numer-
ous studies investigated the interactions between climate change and
the carbon cycle (e.g., Friedlingstein et al., 2006), methane cycle (e.g.,
O’Connor et al., 2010), ozone (Cionni et al., 2011) or aerosols (e.g.,
Carslaw et al., 2010). Many CMIP5 ESMs now include a representa-
tion of the carbon cycle as well as atmospheric chemistry, allowing
interactive projections of GHGs (mainly CO
2
and O
3
) and aerosols. With
such models, projections account for the imposed changes in anthro-
pogenic emissions, but also for changes in natural sources and sinks
as they respond to changes in climate and atmospheric composition. If
included in ESMs, the impact on projected concentration, RF and hence
on climate can be quantified. Climate-induced changes on the carbon
cycle are assessed below, while changes in natural emissions of CH
4
are assessed in Chapter 6, changes in atmospheric chemistry in Chap-
ter 11, and climate–aerosol interactions are assessed in Chapter 7.
12.4.8.1 Carbon Dioxide
As presented in Section 12.3, the CMIP5 experimental design includes,
for the RCP8.5 scenario, experiments driven either by prescribed
anthropogenic CO
2
emissions or concentration. The historical and
21st century emission-driven simulations allow evaluating the cli-
mate response of the Earth system when atmospheric CO
2
and the cli-
mate response are interactively being calculated by the ESMs. In such
ESMs, the atmospheric CO
2
is calculated as the difference between
the imposed anthropogenic emissions and the sum of land and ocean
carbon uptakes. As most of these ESMs account for land use changes
and their CO
2
emissions, the only external forcing is fossil fuel CO
2
emissions (along with all non-CO
2
forcings as in the C-driven RCP8.5
simulations). For a given ESM, the emission driven and concentration
driven simulations would show different climate projections if the
simulated atmospheric CO
2
in the emission driven run is significantly
different from the one prescribed for the concentration driven runs.
This would happen if the ESMs carbon cycle is different from the one
simulated by MAGICC6, the model used to calculate the CMIP5 GHGs
concentrations from the emissions for the four RCPs (Meinshausen et
al., 2011c). When driven by CO
2
concentration, the ESMs can calculate
the fossil fuel CO
2
emissions that would be compatible with the pre-
scribed atmospheric CO
2
trajectory, allowing comparison with the set
of CO
2
emissions initially estimated by the IAMs (Arora et al., 2011;
Jones et al., 2013) (see Section 6.4.3, Box 6.4).
Figure 12.36 shows the simulated atmospheric CO
2
and global aver-
age surface air temperature warming (relative to the 1986–2005 ref-
erence period) for the RCP8.5 emission driven simulations from the
CMIP5 ESMs, compared to the concentration driven simulations from
the same models. Most (seven out of eleven) of the models estimate a
larger CO
2
concentration than the prescribed one. By 2100, the multi-
model average CO
2
concentration is 985 ± 97 ppm (full range 794
to 1142 ppm), while the CO
2
concentration prescribed for the RCP8.5
is 936 ppm. Figure 12.36 also shows the range of atmospheric CO
2
projections when the MAGICC6 model, used to provide the RCP con-
centrations, is tuned to emulate combinations of climate sensitivity
uncertainty taken from 19 CMIP3 models and carbon cycle feedbacks
uncertainty taken from 10 C
4
MIP models, generating 190 model simu-
lations (Meinshausen et al., 2011c; Meinshausen et al., 2011b). The
emulation of the CMIP3/C
4
MIP models shows for the RCP8.5, a range
of simulated CO
2
concentrations of 794 to 1149 ppm (90% confidence
level), extremely similar to what is obtained with the CMIP5 ESMs,
with atmospheric concentration as high as 1150 ppm by 2100, that is,
more than 200 ppm above the prescribed CO
2
concentration.
Global warming simulated by the E-driven runs show higher upper
ends than when atmospheric CO
2
concentration is prescribed. For the
models assessed here, the global surface temperature change (2081–
2100 average relative to 1986–2005 average) ranges between 2.6°C
and 4.7°C, with a multi-model average of 3.7°C ± 0.7°C for the con-
centration driven simulations, while the emission driven simulations
give a range of 2.5°C to 5.6°C, with a multi-model average of 3.9°C
± 0.9°C, that is, 5% larger than for the concentration driven runs. The
models that simulate the largest CO
2
concentration by 2100 have the
largest warming amplification in the emission driven simulations, with
an additional warming of more than 0.5°C.
The uncertainty on the carbon cycle has been shown to be of com-
parable magnitude to the uncertainty arising from physical climate
processes (Gregory et al., 2009). Huntingford et al. (2009) used a simple
model to characterize the relative role of carbon cycle and climate sen-
sitivity uncertainties in contributing to the range of future temperature
changes, concluding that the range of carbon cycle processes represent
about 40% of the physical feedbacks. Perturbed parameter ensembles
systematically explore land carbon cycle parameter uncertainty and
illustrate that a wide range of carbon cycle responses are consistent
with the same underlying model structures and plausible parameter
ranges (Booth et al., 2012; Lambert et al., 2012). Figure 12.37 shows
how the comparable range of future climate change (SRES A1B) arises
from parametric uncertainty in land carbon cycle and atmospheric
feedbacks. The same ensemble shows that the range of atmospheric
CO
2
in the land carbon cycle ensemble is wider than the full SRES con-
centration range (B1 to A1FI scenario).
The CMIP5 ESMs described above do not include the positive feed-
back arising from the carbon release from high latitudes permafrost
thawing under a warming scenario, which could further increase the
atmospheric CO
2
concentration and the warming. Two recent studies
investigated the climate–permafrost feedback from simulations with
models of intermediate complexity (EMICs) that accounts for a per-
mafrost carbon module (MacDougall et al., 2012; Schneider von Deim-
ling et al., 2012). Burke et al. (2012) also estimated carbon loss from
permafrost, from a diagnostic of the present-day permafrost carbon
store and future soil warming as simulated by CMIP5 models. However,
this last study did not quantify the effect on global temperature. Each
of these studies found that the range of additional warming due to the
permafrost carbon loss is quite large, because of uncertainties in future
high latitude soil warming, amount of carbon stored in permafrost
soils, vulnerability of freshly thawed organic material, the proportion
of soil carbon that might be emitted as carbon dioxide via aerobic
decomposition or as methane via anaerobic decomposition (Schneider
von Deimling et al., 2012). For the RCP8.5, the additional warming
from permafrost ranges between 0.04°C and 0.69°C by 2100 although
1097
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
there is medium confidence in these numbers as are the ones on the
amount of carbon released (see Section 12.5.5.4) (MacDougall et al.,
2012; Schneider von Deimling et al., 2012).
12.4.8.2 Changes in Vegetation Cover
Vegetation cover can also be affected by climate change, with forest
cover potentially being decreasing (e.g., in the tropics) or increasing
(e.g., in high latitudes). In particular, the Amazon forest has been
the subject of several studies, generally agreeing that future climate
change would increase the risk tropical Amazon forest being replaced
by seasonal forest or even savannah (Huntingford et al., 2008; Jones
et al., 2009; Malhi et al., 2009). Increase in atmospheric CO
2
would
partly reduce such risk, through increase in water efficiency under ele-
vated CO
2
(Lapola et al., 2009; Malhi et al., 2009). Recent multi-model
estimates based on different CMIP3 climate scenarios and different
dynamic global vegetation models predict a moderate risk of tropical
forest reduction in South America and even lower risk for African and
Asian tropical forests (see also Section 12.5.5.6) (Gumpenberger et al.,
2010; Huntingford et al., 2013).
200
300
400
500
600
700
800
900
1000
1100
1200
CO
2
concentration (ppm)
a Atmospheric CO
2
concentration
c Atmospheric CO
2
concentration
−1
0
1
2
3
4
5
6
7
b Global mean surface air temperature
d Global mean surface air temperature
1850 1900 1950 2000 2050 2100
200
300
400
500
600
700
800
900
1000
1100
1850 1900 1950 2000 2050 2100
−1
0
1
2
3
4
5
6
Global−Mean Temperature relative to 1986−2005 (°C)
CMIP3 & C4MIP emulation:
CMIP5:
CMIP5:
CMIP3 & C4MIP emulation:
90%
68%
50%
Ranges
Concentration-driven default
90%
68%
50%
Ranges
Concentration-driven default
Emission-driven
Emission-driven
Concentration-driven
C
C
Figure 12.36 | Simulated changes in (a) atmospheric CO
2
concentration and (b) global averaged surface temperature (°C) as calculated by the CMIP5 Earth System Models (ESMs)
for the RCP8.5 scenario when CO
2
emissions are prescribed to the ESMs as external forcing (blue). Also shown (b, in red) is the simulated warming from the same ESMs when directly
forced by atmospheric CO
2
concentration (a, red white line). Panels (c) and (d) show the range of CO
2
concentrations and global average surface temperature change simulated by
the Model for the Assessment of Greenhouse Gas-Induced Climate Change 6 (MAGICC6) simple climate model when emulating the CMIP3 models climate sensitivity range and the
Coupled Climate Carbon Cycle Model Intercomparison Project (C
4
MIP) models carbon cycle feedbacks. The default line in (c) is identical to the one in (a).
Figure 12.37 | Uncertainty in global mean temperature from Met Office Hadley Centre
climate prediction model 3 (HadCM3) results exploring atmospheric physics and ter-
restrial carbon cycle parameter perturbations under the SRES A1B scenario (Murphy et
al., 2004; Booth et al., 2012). Relative uncertainties in the Perturbed Carbon Cycle (PCC,
green plume) and Perturbed Atmospheric Processes (PAP, blue plume) on global mean
anomalies of temperature (relative to the 1986–2005 period). The standard simulations
from the two ensembles, HadCM3 (blue solid) and HadCM3C (green solid) are also
shown. Three bars are shown on the right illustrating the 2100 temperature anomalies
associated with the CMIP3/AR4 ensemble (black) the PAP ensemble (blue) and PCC
ensemble (green). The ranges indicate the full range, 10th to 90th, 25th to 75th and
50th percentiles.
CMIP3
PAP
PCC
1098
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Figure 12.38 | Impact of land use change on surface temperature. LUCID-CMIP5 experiments where six ESMs were forced either with or without land use change beyond 2005
under the RCP8.5 scenario. Left maps of changes in total crop and pasture fraction (%) in the RCP8.5 simulations between 2006 and 2100 as implemented in each ESM.Right maps
show the differences in surface air temperature (averaged over the 2071–2100 period) between the simulations with and without land use change beyond 2005. Only statistically
significant changes (p < 0.05) are shown.
Difference in crop and pasture fraction (%) Change in surface air temperature (°C)
1099
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
ESMs simulations with interactive vegetation confirmed known bio-
physical feedback associated with large-scale changes in vegetation.
In the northern high latitudes, warming-induced vegetation expansion
reduces surface albedo, enhancing the warming over these regions
(Falloon et al., 2012; Port et al., 2012), with potentially larger ampli-
fication due to ocean and sea ice response (Swann et al., 2010). Over
tropical forest, reduction of forest coverage would reduce evapotran-
spiration, also leading to a regional warming (Falloon et al., 2012; Port
et al., 2012).
CMIP5 ESMs also include human induced land cover changes (deforest-
ation, reforestation) affecting the climate system through changes in
land surface physical properties (Hurtt et al., 2011). Future changes
in land cover will have an impact on the climate system through bio-
physical and biogeochemical processes (e.g., Pongratz et al., 2010).
Biophysical processes include changes in surface albedo and changes
in partitioning between latent and sensible heat, while biogeochemi-
cal feedbacks essentially include change in CO
2
sources and sinks but
could potentially also include changes in N
2
O or CH
4
emissions. The bio-
physical response to future land cover changes has been investigated
within the SRES scenarios. Using the SRES A2 2100 land cover, Davin et
al. (2007) simulated a global cooling of 0.14 K relatively to a simulation
with present-day land cover, the cooling being largely driven by change
in albedo. Regional analyses have been performed in order to quantify
the biophysical impact of biofuels plantation generally finding a local
to regional cooling when annual crops are replaced by bioenergy crops,
such as sugar cane (Georgescu et al., 2011; Loarie et al., 2011). How-
ever, some energy crops require nitrogen inputs for their production,
leading inevitably to nitrous oxide (N
2
O) emissions, potentially reduc-
ing the direct cooling effect and the benefit of biofuels as an alterna-
tive to fossil fuel emissions. Such emission estimates are still uncertain,
varying strongly for different crops, management methods, soil types
and reference systems (St. Clair et al., 2008; Smeets et al., 2009).
In the context of the Land-Use and Climate, IDentification of robust
impacts (LUCID) project (Pitman et al., 2009) ESMs performed addi-
tional CMIP5 simulations in order to separate the biophysical from
the biogeochemical effects of land use changes in the RCP scenarios.
The LUCID–CMIP5 experiments were designed to complement RCP8.5
and RCP2.6 simulations of CMIP5, both of which showing an intensi-
fication of land use change over the 21st century. The LUCID–CMIP5
analysis was focussed on a difference in climate and land-atmosphere
fluxes between the average of ensemble of simulations with and with-
out land use changes by the end of 21st century (Brovkin et al., 2013).
Due to different interpretation of land use classes, areas of crops and
pastures were specific for each ESM (Figure 12.38, left). On the global
scale, simulated biophysical effects of land use changes projected in
the CMIP5 experiments with prescribed CO
2
concentrations were not
significant. However, these effects were significant for regions with
land use changes >10%. Only three out of six participating models,
CanESM2, HadGEM2-ES and MIROC-ESM, reveal statistically signifi-
cant changes in regional mean annual mean surface air temperature
for the RCP8.5 scenario (Figure 12.38, right). However, there is low
confidence on the overall effect as there is no agreement among the
models on the sign of the global average temperature change due
to the biophysical effects of land use changes (Brovkin et al., 2013).
Changes in land surface albedo, available energy, latent and sensible
heat fluxes were relatively small but significant in most of ESMs for
regions with substantial land use changes. The scale of climatic effects
reflects a small magnitude of land use changes in both the RCP2.6 and
8.5 scenarios and their limitation mainly to the tropical and subtropical
regions where differences between biophysical effects of forests and
grasslands are less pronounced than in mid- and high latitudes. LUCID-
CMIP5 did not perform similar simulations for the RCP4.5 or RCP6.0
scenarios. As these two scenarios show a global decrease of land use
area, one might expect their climatic impact to be different from the
one seen in the RC2.6 and RCP8.5.
12.4.9 Consistency and Main Differences Between
Coupled Model Intercomparison Project Phase 3/
Coupled Model Intercomparison Project Phase 5
and Special Report on Emission Scenarios/
Representative Concentration Pathways
In the experiments collected under CMIP5, both models and scenario
have changed with respect to CMIP3 making a comparison with earlier
results and the scientific literature they generated (on which some of
this chapter’s content is still based) complex. The set of models used
in AR4 (the CMIP3 models) have been superseded by the new CMIP5
models (Table 12.1; Chapter 9) and the SRES scenarios have been
replaced by four RCPs (Section 12.3.1). In addition, the baseline period
used to compute anomalies has advanced 6 years, from 1980–1999 to
1986–2005.
SRES A1B
2080–2099 - 1980–1999
RCP6.0
2081–2100 - 1986–2005
Global mean temperature anomaly (°C)
4.0
3.5
3.0
2.5
2.0
1.5
1.0
CMIP3 CMIP5 (em)
CMIP5+CMIP3 (em) CMIP5
Figure 12.39 | Global mean temperature anomalies at the end of the 21st century
from General Circulation Model (GCM) experiments and emulators comparing CMIP3/
CMIP5 responses under SRES A1B and RCP6.0. The boxes and whiskers indicate the
5th percentile, mean value – 1 standard deviation, mean, mean value + 1 standard
deviation and 95th percentile of the distributions. The first box-and-whiskers on the
left is computed directly from the CMIP3 ensemble and corresponds to the numbers
quoted in AR4. The emulated SRES A1B projections (second from left) of CMIP5 are
obtained by the method of Good et al. (2011a) and are calculated for the period 2080-
2099 expressed with respect to the AR4 baseline period of 1980–1999. Because of the
method, the subset of CMIP5 that are emulated are restricted to those with pre-indus-
trial control, abrupt 4 × CO
2
, historical, RCP4.5 and RCP8.5 simulations. The emulated
RCP6.0 projections of CMIP3 (third from left, see also Figure 12.8) are from Knutti and
Sedláček (2013) obtained using the method of Meinshausen et al. (2011b; 2011c) and
are calculated for the slightly different future period 2081–2100 to be consistent with
the rest of this chapter, and are expressed with respect to the AR5 baseline period of
1986–2005. The box-and-whiskers fourth from the left are a graphical representation of
the numbers shown in Table 12.2. The final box-and-whiskers on the right is a combina-
tion of CMIP5 model output and emulation of CMIP5 RCP6.0 numbers for those models
that did not run RCP6.0.
1100
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
It would be extremely costly computationally to rerun the full CMIP3
ensemble under the new RCPs and/or the full CMIP5 ensemble under
the old SRES scenarios in order to separate model and scenario effects.
In the absence of a direct comparison, we rely on simplified model-
ling frameworks to emulate CMIP3/5 SRES/RCP model behaviour and
compare them. Figure 12.39 shows an emulation of the global mean
temperature response at the end of the 21st century that one would
expect from the CMIP5 models if they were run under SRES A1B. In this
case, anomalies are computed with respect to 1980–1999 for direct
comparison with the values reported in AR4 (Meehl et al., 2007b)
which used that baseline. The method used to emulate the SRES A1B
response of the CMIP5 is documented by Good et al. (2011a; 2013).
Ensemble-mean A1B RF was computed from CMIP3 projections using
the Forster and Taylor (2006) method, scaled to ensure consistency
with the forcing required by the method. The simple model is only used
to predict the temperature difference between A1B and RCP8.5, and
between A1B and RCP4.5 separately for each model. These differenc-
es are then added to CMIP5 GCM simulations of RCP8.5 and RCP4.5
respectively, and averaged to give a single A1B estimate. The emulated
CMIP5 SRES A1B results show a slightly larger mean response than the
actual CMIP3 models, with a similar spread (±1 standard deviation is
used in this case). The main reason for this is the slightly larger mean
transient climate response (TCR) in the subset of CMIP5 models avail-
able in comparison with the AR4 CMIP3 models. An alternative emula-
tion is presented by Knutti and Sedláček (2013) who use the simplified
MAGICC models with parameters chosen to emulate the response of
the CMIP3 models to RCP6.0 forcing, with anomalies expressed with
respect to the 1986–2005 baseline period (Figure 12.39). They too find
a larger mean response in the CMIP5 case but also a larger spread (±1
standard deviation) in CMIP5. Uncertainties in the different approach-
es to emulating climate model simulations, for example estimating the
non-GHG RF, and the small sample sizes of CMIP3 and CMIP5 make
it difficult to draw conclusions on the statistical significance of the
differences displayed in Figure 12.39, but the same uncertainties lead
us to conclude that on the basis of these analyses there appears to
be no fundamental difference between the behaviour of the CMIP5
ensemble, in comparison with CMIP3.
Meinshausen et al. (2011a; 2011b) tuned MAGICC6 to emulate 19
GCMs from CMIP3. The results are temperature projections and their
uncertainties (based on the empirical distribution of the ensemble)
under each of the RCPs, extended to year 2500 (under constant emis-
sions for the lowest RCP and constant concentrations for the remain-
ing three). In the same paper, an ensemble produced by combining
carbon cycle parameter calibration to nine C
4
MIP models with the 19
CMIP3 model parameter calibrations is also used to estimate the emis-
sions implied by the various concentration pathways, had the CMIP3
models included a carbon cycle component. Rogelj et al. (2012) used
the same tool but performed a fully probabilistic analysis of the SRES
and RCP scenarios using a parameter space that is consistent with
SRESB1
SRESA1T
SRESB2
SRESA1B
SRESA2
SRESA1FI
RCP3-PD
RCP4.5
RCP6
RCP8.5
Temperature increase in 2090-2099 relative to 1980-1999 (°C)
1950 2000 2050 2100 2150 2200 2250 2300
0
1
2
3
4
5
6
7
8
0
1
2
3
4
5
6
7
8
Temperature increase relative to 1980-1999 (°C)
Temperature increase in 2090-2099 relative to pre-industrial (°C)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
Temperature increase relative to pre-industrial (°C)
SRES scenarios RCPs
ab
likely range (-40 to +60% around mean)
best estimate
median
66% range
90% range
median
90% range for emission-driven RCPs
66% range for emission-driven RCPs
IPCC AR4 values
Rogelj et al. (2012)
RCP3-PD
RCP4.5
RCP6
1980-1999 period
2090-2099 period
RCP8.5
Figure 12.40 | Temperature projections for SRES scenarios and the RCPs. (a) Time-evolving temperature distributions (66% range) for the four RCP scenarios computed with
the ECS distribution from Rogelj et al. (2012) and a model setup representing closely the carbon-cycle and climate system uncertainty estimates of the AR4 (grey areas). Median
paths are drawn in yellow. Red shaded areas indicate time periods referred to in panel b. (b) Ranges of estimated average temperature increase between 2090 and 2099 for SRES
scenarios and the RCPs respectively. Note that results are given both relative to 1980–1999 (left scale) and relative to pre-industrial (right scale). Yellow ranges indicate results
obtained by Rogelj et al. (2012). Colour-coding of AR4 ranges is chosen to be consistent with AR4 (Meehl et al., 2007b). RCP2.6 is labelled as RCP3-PD here.
1101
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
CMIP3/C
4
MIP but a more general uncertainty characterization for key
quantities like equilibrium climate sensitivity, similarly to the approach
utilized by Meinshausen et al. (2009). Observational or other historical
constraints are also used in this study and the analysis is consistent
with the overall assessment of sources and ranges of uncertainties for
relevant quantities (equilibrium climate sensitivity above all) from AR4
(Meehl et al., 2007b , Box 10.2). Figure 12.40 summarizes results of this
probabilistic comparison for global temperature. The RCPs span a large
range of stabilization, mitigation and non-mitigation pathways and
the resulting range of temperature changes are larger than those pro-
duced under SRES scenarios, which do not consider mitigation options.
The SRES results span an interval between just above 1.0°C and 6.5°C
when considering the respective likely ranges of all scenarios, including
B1 as the lowest and A1FI as the highest. Emissions under RCP8.5 are
highest and the resulting temperature changes likely range from 4.0°C
to 6.1°C by 2100. The lowest RCP2.6 assumes significant mitigation
and the global temperature change likely remains below 2°C.
Similar temperature change projections by the end of the 21st century
are obtained under RCP8.5 and SRES A1FI, RCP6 and SRES B2 and
RCP4.5 and SRES B1. There remain large differences though in the tran-
sient trajectories, with rates of change slower or faster for the different
pairs. These differences can be traced back to the interplay of the (neg-
ative) short-term effect of sulphate aerosols and the (positive) effect of
long-lived GHGs. Impact studies may be sensitive to the differences in
these temporal profiles so care should be taken in approximating SRES
with RCPs and vice versa.
While simple models can separate the effect of the scenarios and the
model response, no studies are currently available that allow an attri-
bution of the CMIP3-CMIP5 differences to changes in the transient
climate response, the carbon cycle, and the inclusion of new processes
(chemistry, land surface, vegetation). The fact that these sets of CMIP3
and CMIP5 experiments do not include emission-driven runs would
suggest that differences in the representation of the carbon cycle are
very unlikely to explain differences in the simulations, since the only
Figure 12.41 | Patterns of temperature (left column) and percent precipitation change (right column) for the CMIP3 models average (first row) and CMIP5 models average (second
row), scaled by the corresponding global average temperature changes. The patterns are computed in both cases by taking the difference between the averages over the last 20
years of the 21st century experiments (2080–2099 for CMIP3 and 2081–2100 for CMIP5) and the last twenty years of the historic experiments (1980–1999 for CMIP3, 1986–2005
for CMIP5) and rescaling each difference by the corresponding change in global average temperature. This is done first for each individual model, and then the results are averaged
across models. For the CMIP5 patterns, the RCP2.6 simulation of the FIO-ESM model was excluded because it did not show any warming by the end of the 21st century, thus not
complying with the method requirement that the pattern be estimated at a time when the temperature change signal from CO
2
increase has emerged. Stippling indicates a measure
of significance of the difference between the two corresponding patterns obtained by a bootstrap exercise. Two subsets of the pooled set of CMIP3 and CMIP5 ensemble members
of the same size as the original ensembles, but without distinguishing CMIP3 from CMIP5 members, were randomly sampled 500 times. For each random sample we compute the
corresponding patterns and their difference, then the true difference is compared, grid-point by grid-point, to the distribution of the bootstrapped differences, and only grid-points
at which the value of the difference falls in the tails of the bootstrapped distribution (less than the 2.5 percentiles or the 97.5 percentiles) are stippled.
1102
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
effect of changes in the carbon cycle representation would affect the
land surface, and thus would have only a minor effect on the climate
response at the global scale.
Figure 12.41 shows a comparison of the patterns of warming and
precipitation change from CMIP3 (using 23 models and three SRES
scenarios) and CMIP5 (using 46 models and four RCPs), utilizing the
pattern scaling methodology (Section 12.4.2). The geographic patterns
of mean change are very similar across the two ensembles of models,
with pattern correlations of 0.98 for temperature and 0.90 for precipi-
tation changes. However there exist significant differences in the abso-
lute values of the patterns, if not in their geographic shapes. A simple
bootstrapping exercise that pooled together all models and scenari-
os and resampled 500 times the same numbers of models/scenarios
divided into two groups, but without distinguishing CMIP3 from CMIP5
(and thus SRES from RCPs) allows to compute a measure of signifi-
cance of the actual differences in the patterns. Stippling in Figure 12.41
marks the large regions where the difference is significant for temper-
ature and precipitation patterns. The temperature pattern from CMIP5
shows significantly larger warming per degree Celsius of global mean
temperature change in the NH and less warming per degree Celsius in
the SH compared to the corresponding pattern from CMIP3. For precip-
itation patterns, CMIP5 shows significantly larger increases per degree
Celsius in the NH and significantly larger decreases per degree Celsius
in the SH compared to CMIP3. Even in this case we do not have studies
that allow tracing the source of these differences to specific changes in
models’ configurations, processes represented or scenarios run.
Knutti and Sedláček (2013) attempt to identify or rule out at least
some of these sources. Differences in model projections spread or its
counterpart, robustness, between CMIP3 and CMIP5 are discussed,
and it is shown that by comparing the behaviour of only a subset
of 11 models, contributed to the two CMIPs by the same group of
institutions, the robustness of CMIP5 versus that of CMIP3 actually
decreases slightly. This would suggest that the enhanced robustness
of CMIP5 is not clearly attributable to advances in modelling, and may
be a result of the fact that the CMIP5 ensemble contains different
versions of the same model that are counted as independent in this
measure of robustness.
A comparison of CMIP3 and CMIP5 results for extreme indices is pro-
vided in Sections 12.4.3.3 and Figure 12.13 for temperature extremes,
and Section 12.4.5.5 and Figure 12.26 for extremes in the water cycle.
12.5 Climate Change Beyond 2100,
Commitment, Stabilization and
Irreversibility
This section discusses the long term (century to millennia) climate
change based on the RCP scenario extensions and idealized scenari-
os, the commitment from current atmospheric composition and from
past emissions, the concept of cumulative carbon and the resulting
constraints on emissions for various temperature targets. The term
irreversibility is used in various ways in the literature. This report defines
a perturbed state as irreversible on a given time scale if the recov-
ery time scale from this state due to natural processes is significantly
longer than the time it takes for the system to reach this perturbed
state (see Glossary), for example, the climate change resulting from
the long residence time of a CO
2
perturbation in the atmosphere. These
results are discussed in Sections 12.5.2 to 12.5.4. Aspects of irreversi-
bility in the context of abrupt change, multiple steady states and hys-
teresis are discussed in Section 12.5.5 and in Chapter 13 for ice sheets
and sea level rise.
12.5.1 Representative Concentration Pathway Extensions
The CMIP5 intercomparison project includes simulations extending the
four RCP scenarios to the year 2300 (see Section 12.3.1). This allows
exploring the longer-term climate response to idealized GHG and aer-
osols forcings (Meinshausen et al., 2011c). Continuing GHG emissions
beyond 2100 as in the RCP8.5 extension induces a total RF above 12
W m
–2
by 2300, while sustaining negative emissions beyond 2100, as
in the RCP2.6 extension, induces a total RF below 2 W m
–2
by 2300.
The projected warming for 2281–2300, relative to 1986–2005, is 0.6°C
(range 0.0°C to 1.2°C) for RCP2.6, 2.5°C (range 1.5°C to 3.5°C) for
RCP4.5, and 7.8°C (range 3.0°C to 12.6°C) for RCP8.5 (medium confi-
dence, based on a limited number of CMIP5 simulations) (Figures 12.3
and 12.5, Table 12.2).
EMICs simulations have been performed following the same CMIP5
protocol for the historical simulation and RCP scenarios extended
to 2300 (Zickfeld et al., 2013). These scenarios have been prolonged
beyond 2300 to investigate longer-term commitment and irreversibility
(see below). Up to 2300, projected warming and the reduction of the
AMOC as simulated by the EMICs are similar to those simulated by the
CMIP5 ESMs (Figures 12.5 and 12.42).
12.5.2 Climate Change Commitment
Climate change commitment, the fact that the climate will change
further after the forcing or emissions have been eliminated or held
constant, has attracted increased attention by scientists and poli-
cymakers shortly before the completion of IPCC AR4 (Hansen et al.,
2005a; Meehl et al., 2005b, 2006; Wigley, 2005) (see also AR4 Section
10.7.1). However, the argument that the surface response would lag
the RF due to the large thermal reservoir of the ocean in fact goes back
much longer (Bryan et al., 1982; Hansen et al., 1984, 1985; Siegenthal-
er and Oeschger, 1984; Schlesinger, 1986; Mitchell et al., 2000; Weth-
erald et al., 2001). The discussion in this section is framed largely in
terms of temperature change, but other changes in the climate system
(e.g., precipitation) are closely related to changes in temperature (see
Sections 12.4.1.1 and 12.4.2). A summary of how past emissions relate
to future warming is also given in FAQ 12.3.
The Earth system has multiple response time scales related to different
thermal reservoirs (see also Section 12.5.3). For a step change in forcing
(instantaneous increase in the magnitude of the forcing and constant
forcing after that), a large fraction of the total of the surface tempera-
ture response will be realized within years to a few decades (Brasseur
and Roeckner, 2005; Knutti et al., 2008a; Murphy et al., 2009; Hansen et
al., 2011). The remaining response, realized over centuries, is controlled
by the slow mixing of the energy perturbation into the ocean (Stouffer,
2004). The response time scale depends on the amount of ocean mixing
1103
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
and the strength of climate feedbacks, and is longer for higher climate
sensitivity (Hansen et al., 1985; Knutti et al., 2005). The transient cli-
mate response is therefore smaller than the equilibrium response, in
particular for high climate sensitivities. This can also be interpreted as
the ocean heat uptake being a negative feedback (Dufresne and Bony,
2008; Gregory and Forster, 2008). Delayed responses can also occur
due to processes other than ocean warming, for example, vegetation
change (Jones et al., 2009) or ice sheet melt that continues long after
the forcing has been stabilized (see Section 12.5.3).
Several forms of commitment are often discussed in the literature. The
most common is the ‘constant composition commitment’, the warm-
ing that would occur after stabilizing all radiative constituents at a
given year (for example year 2000) levels. For year 2000 commitment,
AOGCMs estimated a most likely value of about 0.6°C for 2100 (rel-
ative to 1980–1999, AR4 Section 10.7.1). A present-day composition
commitment simulation is not part of CMIP5, so direct comparison
with CMIP3 is not possible. However, the available CMIP5 results
based on the RCP4.5 extension with constant RF (see Section 12.5.1)
are consistent with those numbers, with an additional warming of
about 0.5°C 200 years after stabilization of the forcing (Figures 12.5
and 12.42).
500
1000
1500
2000
2100 2150 2200 2250
500
1000
1500
2000
(ppmv)
2000 2050 2300
RCP 2.6
RCP 4.5
RCP 6.0
RCP 8.5
a
0
2
4
6
8
10
0
2
4
6
8
10
(
o
C)
b
Year
(Sv)
c
2100 2150 2200 22502000 2050 2300
Atmospheric CO
2
Surface air temperature change
Change in Atlantic meridional overturning circulation
5
0
-5
-10
-15
Figure 12.42 | (a) Atmospheric CO
2
, (b) projected global mean surface temperature
change and (c) projected change in the Atlantic meridional overturning circulation, as
simulated by EMICs for the four RCPs up to 2300 (Zickfeld et al., 2013). A 10-year
smoothing was applied. Shadings and bars denote the minimum to maximum range.
The dashed line on (a) indicates the pre-industrial CO
2
concentration.
A measure of constant composition commitment is the fraction of real-
ized warming which can be estimated as the ratio of the warming at a
given time to the long-term equilibrium warming (e.g., Stouffer, 2004;
Meehl et al., 2007b, Section 10.7.2; Eby et al., 2009; Solomon et al.,
2009). EMIC simulations have been performed with RCPs forcing up to
2300 prolonged until the end of the millennium with a constant forc-
ing set at the value reached by 2300 (Figure 12.43). When the forcing
stabilizes, the fraction of realized warming is significantly below unity.
However, the fraction of realized warming depends on the history of
the forcing. For the RCP4.5 and RCP6.0 extension scenarios with early
stabilization, it is about 75% at the time of forcing stabilization; while
for RCP8.5, with stabilization occurring later, it is about 85% (see Figure
12.43); but for a 1% yr
–1
CO
2
increase to 2 × CO
2
or 4 × CO
2
and con-
stant forcing thereafter, the fraction of realized warming is much small-
er, about 40 to 70% at the time when the forcing is kept constant. The
fraction of realized warming rises typically by 10% over the century
following the stabilization of forcing. Due to the long time scales in the
deep ocean, full equilibrium is reached only after hundreds to thou-
sands of years (Hansen et al., 1985; Gregory et al., 2004; Stouffer, 2004;
Meehl et al., 2007b, Section 10.7.2; Knutti et al., 2008a; Danabasoglu
and Gent, 2009; Held et al., 2010; Hansen et al., 2011; Li et al., 2013a).
500
1000
1500
2000
2000 2200 2400 2600 2800 3000
500
1000
1500
2000
(ppmv)
RCP 2.6
RCP 4.5
RCP 6.0
RCP 8.5
a
0
2
4
6
8
10
(
o
C)
0
2
4
6
8
10
b
Year
Fraction of realized warming
Year
c
Atmospheric CO
2
Surface air temperature change
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
2000 2200 2400 2600 2800 3000
Figure 12.43 | (a) Atmospheric CO
2
, (b) projected global mean surface temperature
change and (c) fraction of realized warming calculated as the ratio of global tempera-
ture change at a given time to the change averaged over the 2980–2999 time period,
as simulated by Earth System Models of Intermediate Complexity (EMICs) for the 4
RCPs up to 2300 followed by a constant (year 2300 level) radiative forcing up to the
year 3000 (Zickfeld et al., 2013). A 10-year smoothing was applied. Shadings and bars
denote the minimum to maximum range. The dashed line on (a) indicates the pre-indus-
trial CO
2
concentration.
1104
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
‘Constant emission commitment’ is the warming that would result
from maintaining annual anthropogenic emissions at the current level.
Few studies exist but it is estimated to be about 1°C to 2.5°C by 2100
assuming constant (year 2010) emissions in the future, based on the
MAGICC model calibrated to CMIP3 and C
4
MIP models (Meinshausen
et al., 2011a; Meinshausen et al., 2011b) (see FAQ 12.3). Such a scenar-
io is different from non-intervention economic scenarios, and it does not
stabilize global temperature, as any plausible emission path after 2100
would cause further warming. It is also different from a constant cumu-
lative emission scenario which implies zero emissions in the future.
Another form of commitment involves climate change when anthropo-
genic emissions are set to zero (‘zero emission commitment’). Results
from a variety of models ranging from EMICs (Meehl et al., 2007b;
Weaver et al., 2007; Matthews and Caldeira, 2008; Plattner et al.,
2008; Eby et al., 2009; Solomon et al., 2009; Friedlingstein et al., 2011)
to ESMs (Frölicher and Joos, 2010; Gillett et al., 2011; Gillett et al.,
2013) show that abruptly setting CO
2
emissions to zero (keeping other
forcings constant if accounted for) results in approximately constant
global temperature for several centuries onward. Those results indicate
that past emissions commit us to persistent warming for hundreds of
years, continuing at about the level of warming that has been realized.
On near equilibrium time scales of a few centuries to about a mil-
lennium, the temperature response to CO
2
emissions is controlled by
climate sensitivity (see Box 12.2) and the cumulative airborne fraction
of CO
2
over these time scales. After about a thousand years (i.e., near
thermal equilibrium) and cumulative CO
2
emissions less than about
2000 PgC, approximately 20 to 30% of the cumulative anthropogenic
carbon emissions still remain in the atmosphere (Montenegro et al.,
2007; Plattner et al., 2008; Archer et al., 2009; Frölicher and Joos, 2010;
Joos et al., 2013) (see Box 6.1) and maintain a substantial temperature
response long after emissions have ceased (Friedlingstein and Solo-
mon, 2005; Hare and Meinshausen, 2006; Weaver et al., 2007; Mat-
thews and Caldeira, 2008; Plattner et al., 2008; Eby et al., 2009; Lowe et
al., 2009; Solomon et al., 2009, 2010; Frölicher and Joos, 2010; Zickfeld
et al., 2012). In the transient phase, on a 100- to 1000-year time scale,
the approximately constant temperature results from a compensation
between delayed commitment warming (Meehl et al., 2005b; Wigley,
2005) and the reduction in atmospheric CO
2
resulting from ocean and
land carbon uptake as well as from the nonlinear dependence of RF on
atmospheric CO
2
(Meehl et al., 2007b; Plattner et al., 2008; Solomon
et al., 2009; Solomon et al., 2010). The commitment associated with
past emissions depends, as mentioned above, on the value of climate
sensitivity and cumulative CO
2
airborne fraction, but it also depends on
the choices made for other RF constituents. In a CO
2
only case and for
equilibrium climate sensitivities near 3°C, the warming commitment
(i.e., the warming relative to the time when emissions are stopped)
is near zero or slightly negative. For high climate sensitivities, and in
particular if aerosol emissions are eliminated at the same time, the
commitment from past emission can be significantly positive, and is
a superposition of a fast response to reduced aerosols emissions and
a slow response associated with high climate sensitivities (Brasseur
and Roeckner, 2005; Hare and Meinshausen, 2006; Armour and Roe,
2011; Knutti and Plattner, 2012; Matthews and Zickfeld, 2012) (see
FAQ 12.3). In the real world, the emissions of CO
2
and non-CO
2
forcing
agents are of course coupled. All of the above studies support the con-
clusion that temperatures would decrease only very slowly (if at all),
even for strong reductions or complete elimination of CO
2
emissions,
and might even increase temporarily for an abrupt reduction of the
short-lived aerosols (FAQ 12.3). The implications of this fact for climate
stabilization are discussed in Section 12.5.4.
New EMIC simulations with pre-industrial CO
2
emissions and zero
non-CO
2
forcings after 2300 (Zickfeld et al., 2013) confirm this behav-
iour (Figure 12.44) seen in many earlier studies (see above). Switching
off anthropogenic CO
2
emissions in 2300 leads to a continuous slow
decline of atmospheric CO
2
, to a significantly slower decline of global
temperature and to a continuous increase in ocean thermal expansion
Figure 12.44 | (a) Compatible anthropogenic CO
2
emissions up to 2300, followed by
zero emissions after 2300, (b) prescribed atmospheric CO
2
concentration up to 2300
followed by projected CO
2
concentration after 2300, (c) global mean surface tempera-
ture change and (d) ocean thermal expansion as simulated by Earth System Models of
Intermediate Complexity (EMICs) for the four concentration driven RCPs with all forcings
included (Zickfeld et al., 2013). A 10-year smoothing was applied. The drop in tempera-
ture in 2300 is a result of eliminating all non-CO
2
forcings along with CO
2
emissions.
Shadings and bars denote the minimum to maximum range. The dashed line on (b)
indicates the pre-industrial CO
2
concentration.
1105
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
over the course of the millennium. Larger forcings induce longer delays
before the Earth system would reach equilibrium. For RCP8.5, by year
3000 (700 years after emissions have ceased) global temperature has
decreased only by 1°C to 2°C (relative to its peak value by 2300) and
ocean thermal expansion has almost doubled (relative to 2300) and is
still increasing (Zickfeld et al., 2013).
The previous paragraph discussed climate change commitment from
GHGs that have already been emitted. Another form of commitment
refers to climate change associated with heat and carbon that has
gone into the land surface and oceans. This would be relevant to the
consequences of a one-time removal of all of the excess CO
2
in the
atmosphere and is computed by taking a transient simulation and
instantaneously setting atmospheric CO
2
concentrations to initial
(pre-industrial) values (Cao and Caldeira, 2010). In such an extreme
case, there would be a net flux of CO
2
from the ocean and land surface
to the atmosphere, releasing an amount of CO
2
representing about
30% of what was removed from the atmosphere, i.e., the airborne frac-
tion applies equally to positive and negative emissions, and it depends
on the emissions history. A related form of experiment investigates
the consequences of an initial complete removal followed by sustained
removal of any CO
2
returned to the atmosphere from the land sur-
face and oceans, and is computed by setting atmospheric CO
2
con-
centrations to pre-industrial values and maintaining this concentration
(Cao and Caldeira, 2010). In this case, only about one-tenth of the
pre-existing temperature perturbation persists for more than half of a
century. A similar study performed with a GFDL AOGCM where forcing
was instantaneously returned to its pre-industrial value, found larger
residual warming, up to 30% of the pre-existing warming (Held et al.,
2010).
Several studies on commitment to past emissions have demonstrat-
ed that the persistence of warming is substantially longer than the
lifetime of anthropogenic GHGs themselves, as a result of nonlinear
absorption effects as well as the slow heat transfer into and out of
the ocean. In much the same way as the warming to a step increase of
forcing is delayed, the cooling after setting RF to zero is also delayed.
Loss of excess heat from the ocean will lead to a positive surface air
temperature anomaly for decades to centuries (Held et al., 2010; Solo-
mon et al., 2010; Bouttes et al., 2013).
A more general form of commitment is the question of how much
warming we are committed to as a result of inertia and hence com-
mitments related to the time scales for energy system transitions and
other societal, economic and technological aspects (Grubb, 1997;
Washington et al., 2009; Davis et al., 2010). For example, Davis et al.
(2010) estimated climate commitment of 1.3°C (range 1.1°C to 1.4°C,
relative to pre-industrial) from existing CO
2
-emitting devices under
specific assumptions regarding their lifetimes. These forms of commit-
ment, however, are strongly based on political, economic and social
assumptions that are outside the domain of IPCC WGI and are not
further considered here.
12.5.3 Forcing and Response, Time Scales of Feedbacks
Equilibrium climate sensitivity (ECS), transient climate response
(TCR) and climate feedbacks are useful concepts to characterize the
response of a model to an external forcing perturbation. However,
there are limitations to the concept of RF (Joshi et al., 2003; Shine et
al., 2003; Hansen et al., 2005b; Stuber et al., 2005), and the separation
of forcings and fast (or rapid) responses (e.g., clouds changing almost
instantaneously as a result of CO
2
-induced heating rates rather than
as a response to the slower surface warming) is sometimes difficult
(Andrews and Forster, 2008; Gregory and Webb, 2008). Equilibrium
warming also depends on the type of forcing (Stott et al., 2003; Hansen
et al., 2005b; Davin et al., 2007). ECS is time or state dependent in
some models (Senior and Mitchell, 2000; Gregory et al., 2004; Boer et
al., 2005; Williams et al., 2008; Colman and McAvaney, 2009; Colman
and Power, 2010), and in some but not all models climate sensitivity
from a slab ocean version differs from that of coupled models or the
effective climate sensitivity (see Glossary) diagnosed from a transient
coupled integration (Gregory et al., 2004; Danabasoglu and Gent,
2009; Li et al., 2013a). The computational cost of coupled AOGCMs is
often prohibitively large to run simulations to full equilibrium, and only
a few models have performed those (Manabe and Stouffer, 1994; Voss
and Mikolajewicz, 2001; Gregory et al., 2004; Danabasoglu and Gent,
2009; Li et al., 2013a). Because of the time dependence of effective
climate sensitivity, fitting simple models to AOGCMs over the first few
centuries may lead to errors when inferring the response on multi-cen-
tury time scales. In the HadCM3 case the long-term warming would be
underestimated by 30% if extrapolated from the first century (Gregory
et al., 2004), in other models the warming of the slab and coupled
model is almost identical (Danabasoglu and Gent, 2009). The assump-
tion that the response to different forcings is approximately additive
appears to be justified for large-scale temperature changes but limited
for other climate variables (Boer and Yu, 2003; Sexton et al., 2003; Gil-
lett et al., 2004; Meehl et al., 2004; Jones et al., 2007). A more complete
discussion of the concept of ECS and the limitations is given in Knutti
and Hegerl (2008). The CMIP5 model estimates of ECS and TCR are
also discussed in Section 9.7. Despite all limitations, the ECS and TCR
remain key concepts to characterize the transient and near equilibrium
warming as a response to RF on time scales of centuries. Their overall
assessment is given in Box 12.2.
A number of recent studies suggest that equilibrium climate sensitiv-
ities determined from AOGCMs and recent warming trends may sig-
nificantly underestimate the true Earth system sensitivity (see Glossa-
ry) which is realized when equilibration is reached on millennial time
scales (Hansen et al., 2008; Rohling et al., 2009; Lunt et al., 2010; Pagani
et al., 2010; Rohling and Members, 2012). The argument is that slow
feedbacks associated with vegetation changes and ice sheets have
their own intrinsic long time scales and are not represented in most
models (Jones et al., 2009). Additional feedbacks are mostly thought
to be positive but negative feedbacks of smaller magnitude are also
simulated (Swingedouw et al., 2008; Goelzer et al., 2011). The climate
sensitivity of a model may therefore not reflect the sensitivity of the
full Earth system because those feedback processes are not considered
(see also Sections 10.8, 5.3.1 and 5.3.3.2; Box 5.1). Feedbacks deter-
mined in very different base state (e.g., the Last Glacial Maximum)
differ from those in the current warm period (Rohling and Members,
2012), and relationships between observables and climate sensitiv-
ity are model dependent (Crucifix, 2006; Schneider von Deimling et
al., 2006; Edwards et al., 2007; Hargreaves et al., 2007, 2012). Esti-
mates of climate sensitivity based on paleoclimate archives (Hansen
1106
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Frequently Asked Questions
FAQ 12.3 | What Would Happen to Future Climate if We Stopped Emissions Today?
Stopping emissions today is a scenario that is not plausible, but it is one of several idealized cases that provide
insight into the response of the climate system and carbon cycle. As a result of the multiple time scales in the climate
system, the relation between change in emissions and climate response is quite complex, with some changes still
occurring long after emissions ceased. Models and process understanding show that as a result of the large ocean
inertia and the long lifetime of many greenhouse gases, primarily carbon dioxide, much of the warming would
persist for centuries after greenhouse gas emissions have stopped.
When emitted in the atmosphere, greenhouse gases get removed through chemical reactions with other reactive
components or, in the case of carbon dioxide (CO
2
), get exchanged with the ocean and the land. These processes
characterize the lifetime of the gas in the atmosphere, defined as the time it takes for a concentration pulse to
decrease by a factor of e (2.71). How long greenhouse gases and aerosols persist in the atmosphere varies over a
wide range, from days to thousands of years. For example, aerosols have a lifetime of weeks, methane (CH
4
) of
about 10 years, nitrous oxide (N
2
O) of about 100 years and hexafluoroethane (C
2
F
6
) of about 10,000 years. CO
2
is
more complicated as it is removed from the atmosphere through multiple physical and biogeochemical processes in
the ocean and the land; all operating at different time scales. For an emission pulse of about 1000 PgC, about half
is removed within a few decades, but the remaining fraction stays in the atmosphere for much longer. About 15 to
40% of the CO
2
pulse is still in the atmosphere after 1000 years.
As a result of the significant lifetimes of major anthropogenic greenhouse gases, the increased atmospheric concen-
tration due to past emissions will persist long after emissions are ceased. Concentration of greenhouse gases would
not return immediately to their pre-industrial levels if emissions were halted. Methane concentration would return
to values close to pre-industrial level in about 50 years, N
2
O concentrations would need several centuries, while
CO
2
would essentially never come back to its pre-industrial level on time scales relevant for our society. Changes
in emissions of short-lived species like aerosols on the other hand would result in nearly instantaneous changes in
their concentrations.
The climate system response to the greenhouse gases
and aerosols forcing is characterized by an inertia,
driven mainly by the ocean. The ocean has a very large
capacity of absorbing heat and a slow mixing between
the surface and the deep ocean. This means that it will
take several centuries for the whole ocean to warm up
and to reach equilibrium with the altered radiative forc-
ing. The surface ocean (and hence the continents) will
continue to warm until it reaches a surface temperature
in equilibrium with this new radiative forcing. The AR4
showed that if concentration of greenhouse gases were
held constant at present day level, the Earth surface
would still continue to warm by about 0.6°C over the
21st century relative to the year 2000. This is the climate
commitment to current concentrations (or constant
composition commitment), shown in grey in FAQ 12.3,
Figure 1. Constant emissions at current levels would fur-
ther increase the atmospheric concentration and result
in much more warming than observed so far (FAQ 12.3,
Figure 1, red lines).
Even if anthropogenic greenhouses gas emissions were
halted now, the radiative forcing due to these long-
lived greenhouse gases concentrations would only
slowly decrease in the future, at a rate determined
by the lifetime of the gas (see above). Moreover, the
1950 2000 2050 2100 2150
0
1
2
3
4
Global surface warming (°C)
Year
90%
85%
80%
68%
50%
Constant Emissions
Zero Emissions
Constant Forcing
Ensemble Range:
FAQ 12.3, Figure 1 | Projections based on the energy balance carbon
cycle model Model for the Assessment of Greenhouse Gas-Induced Climate
Change (MAGICC) for constant atmospheric composition (constant forcing,
grey), constant emissions (red) and zero future emissions (blue) starting in
2010, with estimates of uncertainty. Figure adapted from Hare and Mein-
shausen (2006) based on the calibration of a simple carbon cycle climate
model to all Coupled Model Intercomparison Project Phase 3 (CMIP3) and
Coupled Climate Carbon Cycle Model Intercomparison Project (C4MIP)
models (Meinshausen et al., 2011a; Meinshausen et al., 2011b). Results are
based on a full transient simulation starting from pre-industrial and using
all radiative forcing components. The thin black line and shading denote the
observed warming and uncertainty.
(continued on next page)
1107
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
FAQ 12.3 (continued)
climate response of the Earth System to that radiative forcing would be even slower. Global temperature would
not respond quickly to the greenhouse gas concentration changes. Eliminating CO
2
emissions only would lead to
near constant temperature for many centuries. Eliminating short-lived negative forcings from sulphate aerosols at
the same time (e.g., by air pollution reduction measures) would cause a temporary warming of a few tenths of a
degree, as shown in blue in FAQ 12.3, Figure 1. Setting all emissions to zero would therefore, after a short warming,
lead to a near stabilization of the climate for multiple centuries. This is called the commitment from past emissions
(or zero future emission commitment). The concentration of GHG would decrease and hence the radiative forcing
as well, but the inertia of the climate system would delay the temperature response.
As a consequence of the large inertia in the climate and carbon cycle, the long-term global temperature is largely
controlled by total CO
2
emissions that have accumulated over time, irrespective of the time when they were emit-
ted. Limiting global warming below a given level (e.g., 2°C above pre-industrial) therefore implies a given budget
of CO
2
, that is, higher emissions earlier implies stronger reductions later. A higher climate target allows for a higher
CO
2
concentration peak, and hence larger cumulative CO
2
emissions (e.g., permitting a delay in the necessary emis-
sion reduction).
Global temperature is a useful aggregate number to describe the magnitude of climate change, but not all changes
will scale linearly global temperature. Changes in the water cycle for example also depend on the type of forcing
(e.g., greenhouse gases, aerosols, land use change), slower components of the Earth system such as sea level rise
and ice sheet would take much longer to respond, and there may be critical thresholds or abrupt or irreversible
changes in the climate system.
et al., 2008; Rohling et al., 2009; Lunt et al., 2010; Pagani et al., 2010;
Schmittner et al., 2011; Rohling and Members, 2012), most but not all
based on climate states colder than present, are therefore not neces-
sarily representative for an estimate of climate sensitivity today (see
also Sections 5.3.1, 5.3.3.2, Box 5.1). Also it is uncertain on which time
scale some of those Earth system feedbacks would become significant.
Equilibrium climate sensitivity undoubtedly remains a key quantity,
useful to relate a change in GHGs or other forcings to a global tempera-
ture change. But the above caveats imply that estimates based on past
climate states very different from today, estimates based on time scales
different than those relevant for climate stabilization (e.g., estimates
based on climate response to volcanic eruptions), or based on forcings
other than GHGs (e.g., spatially non-uniform land cover changes, vol-
canic eruptions or solar forcing) may differ from the climate sensitivity
measuring the climate feedbacks of the Earth system today, and this
measure, in turn, may be slightly different from the sensitivity of the
Earth in a much warmer state on time scales of millennia. The TCR and
the transient climate response to cumulative carbon emissions (TCRE)
are often more directly relevant to evaluate short term changes and
emission reductions needed for stabilization (see Section 12.5.4).
12.5.4 Climate Stabilization and Long-term Climate
Targets
This section discusses the relation between emissions and climate
targets, in the context of the uncertainties characterizing both the
transient and the equilibrium climate responses to emissions. ‘Climate
targets’ considered here are both stabilizing temperature at a speci-
fied value and avoiding a warming beyond a predefined threshold.
The latter idea of limiting peak warming is a more general concept
than stabilization of temperature or atmospheric CO
2
, and one that is
more realistic than an exact climate stabilization which would require
perpetual non-zero positive emissions to counteract the otherwise
unavoidable long-term slow decrease in global temperature (Matsuno
et al., 2012a) (Figure 12.44).
12.5.4.1 Background
The concept of stabilization is strongly linked to the ultimate objective
of the UNFCCC, which is ‘to achieve […] stabilization of greenhouse
gas concentrations in the atmosphere at a level that would prevent
dangerous anthropogenic interference with the climate system’. Recent
policy discussions focussed on a global temperature increase, rather
than on GHG concentrations. The most prominent target currently dis-
cussed is the 2°C temperature target, that is, to limit global temper-
ature increase relative to pre-industrial times to below 2°C. The 2°C
target has been used first by the European Union as a policy target in
1996 but can be traced further back (Jaeger and Jaeger, 2010; Randalls,
2010). Climate impacts however are geographically diverse (Joshi et
al., 2011) and sector specific, and no objective threshold defines when
dangerous interference is reached. Some changes may be delayed or
irreversible, and some impacts are likely to be beneficial. It is thus not
possible to define a single critical threshold without value judgments
and without assumptions on how to aggregate current and future
costs and benefits. Targets other than 2°C have been proposed (e.g.,
1.5°C global warming relative to pre-industrial), or targets based on
CO
2
concentration levels, for example, 350 ppm (Hansen et al., 2008).
The rate of change may also be important (e.g., for adaptation). This
section does not advocate or defend any threshold, nor does it judge
1108
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
the economic or political feasibility of such goals, but simply assess-
es the implications of different illustrative climate targets on allowed
carbon emissions, based on our current understanding of climate and
carbon cycle feedbacks.
12.5.4.2 Constraints on Cumulative Carbon Emissions
The current RF from GHGs maintained indefinitely (i.e., the commit-
ment from constant greenhouse gas concentrations) would correspond
to approximately 2°C warming. That, however, does not imply that the
commitment from past emissions has already exceeded 2°C. Part of the
positive RF from GHGs is currently compensated by negative aerosol
forcing, and stopping GHG emissions would lead to a decrease in the
GHG forcing. Actively removing CO
2
from the atmosphere, for example
by the combined use of biomass energy and carbon capture and stor-
age, would further accelerate the decrease in GHG forcing.
The total amount of anthropogenic CO
2
released in the atmosphere
(often termed cumulative carbon emission) is a good indicator of the
atmospheric CO
2
concentration and hence of the global warming
response to CO
2
. The ratio of global temperature change to total cumu-
lative anthropogenic CO
2
emissions (TCRE) is relatively constant over
time and independent of the scenario, but is model dependent as it
depends on the model cumulative airborne fraction of CO
2
and ECS/
TCR (Matthews and Caldeira, 2008; Allen et al., 2009; Gregory et al.,
2009; Matthews et al., 2009; Meinshausen et al., 2009; Zickfeld et al.,
2009; Bowerman et al., 2011; Knutti and Plattner, 2012; Zickfeld et al.,
2012, 2013). This is consistent with an earlier study indicating that
the global warming potential of CO
2
is approximately independent of
the scenario (Caldeira and Kasting, 1993). The concept of a constant
ratio of cumulative emissions of CO
2
to temperature holds well only
until temperatures peak (see Figure 12.45e) and only for smoothly var-
ying cumulative CO
2
emissions (Gillett et al., 2013). It does not hold
for stabilization on millennial time scales or for non-CO
2
forcings, and
there is limited evidence for its applicability for cumulative emissions
exceeding 2000 PgC owing to limited simulations available (Plattner et
al., 2008; Hajima et al., 2012; Matsuno et al., 2012b; Gillett et al., 2013;
Zickfeld et al., 2013). For non-CO
2
forcings with shorter atmospheric
life times than CO
2
the rate of emissions at the time of peak warming
is more important than the cumulative emissions over time (Smith et
al., 2012).
Assuming constant climate sensitivity and fixed carbon cycle feed-
backs, long-term (several centuries to millennium) stabilization of
global temperatures requires eventually the stabilization of atmos-
pheric concentrations (or decreasing concentrations if the temperature
should be stabilized more quickly). This requires decreasing emissions
to near-zero (Jones et al., 2006; Meehl et al., 2007b; Weaver et al.,
2007; Matthews and Caldeira, 2008; Plattner et al., 2008; Allen et al.,
2009; Matthews et al., 2009; Meinshausen et al., 2009; Zickfeld et al.,
2009; Friedlingstein et al., 2011; Gillett et al., 2011; Roeckner et al.,
2011; Knutti and Plattner, 2012; Matsuno et al., 2012a).
The relationships between cumulative emissions and temperature for
various studies are shown in Figure 12.45. Note that some lines mark
the evolution of temperature as a function of emissions over time
while other panels show peak temperatures for different simulations.
Also some models prescribe only CO
2
emissions while others use multi
gas scenarios, and the time horizons differ. The warming is usually
larger if non-CO
2
forcings are considered, since the net effect of the
non-CO
2
forcings is positive in most scenarios (Hajima et al., 2012). Not
all numbers are therefore directly comparable. Matthews et al. (2009)
estimated the TCRE as 1°C to 2.1°C per 1000 PgC (TtC, or 10
12
metric
tonnes of carbon) (5 to 95%) based on the C
4
MIP model range (Figure
12.45a). The ENSEMBLES E1 show a range of 1°C to 4°C per 1000 PgC
(scaled from 0.5°C to 2°C for 500 PgC, Figure 12.45d) (Johns et al.,
2011). Rogelj et al. (2012) estimate a 5 to 95% range of about 1°C to
2°C per 1000 PgC (Figure 12.45e) based on the MAGICC model cali-
brated to the C
4
MIP model range and the likely range of 2°C to 4.5°C
for climate sensitivity given in AR4. Allen et al. (2009) used a simple
model and found 1.3°C to 3.9°C per 1000 PgC (5 to 95%) for peak
warming (Figure 12.45g) and 1.4°C to 2.5°C for TCRE. The EMICs TCRE
simulations suggest a range of about 1.4°C to 2.5°C per 1000 PgC and
a mean of 1.9°C per 1000 PgC (Zickfeld et al., 2013) (Figure 12.45h).
The results of Meinshausen et al. (2009) confirm the approximate lin-
earity between temperature and CO
2
emissions (Figure 12.45b). Their
results are difficult to compare owing to the shorter time period con-
sidered, but the model was found to be consistent with that of Allen et
al. (2009). Zickfeld et al. (2009), using an EMIC, find a best estimate of
about 1.5°C per 1000 PgC. Gillett et al. (2013) find a range of 0.8°C to
2.4°C per 1000 PgC in 15 CMIP5 models and derive an observationally
constrained range of 0.7°C to 2.0°C per 1000 PgC. Results from much
earlier model studies support the near linear relationship of cumulative
emissions and global temperature, even though these studies did not
discuss the linear relationship. An example is given in Figure 12.45c
based on data shown in IPCC TAR Figure 13.3 (IPCC, 2001) and IPCC
AR4 Figure 10.35 (Meehl et al., 2007b). The relationships between
cumulative CO
2
emissions and temperature in CMIP5 are shown in
Figure 12.45f for the 1% yr
–1
CO
2
increase scenarios and in Figure
12.45i for the RCP8.5 emission driven ESM simulations (Gillett et al.,
2013). Compatible emissions from concentration driven CMIP5 ESMs
are discussed in Section 6.4.3.3.
Expert judgement based on the available evidence therefore suggests
that the TCRE is likely between 0.8°C to 2.5°C per 1000 PgC, for cumu-
lative CO
2
emissions less than about 2000 PgC until the time at which
temperature peaks. Under these conditions, and for low to medium
estimates of climate sensitivity, the TCRE is nearly identical to the peak
climate response to cumulative carbon emissions. For high climate
sensitivity, strong carbon cycle climate feedbacks or large cumulative
emissions, the peak warming can be delayed and the peak response
may be different from TCRE, but is often poorly constrained by models
and observations. The range of TCRE assessed here is consistent with
other recent attempts to synthesize the available evidence (NRC, 2011;
Matthews et al., 2012). The results by Schwartz et al. (2010, 2012)
imply a much larger warming for the carbon emitted over the historical
period and have been questioned by Knutti and Plattner (2012) for
neglecting the relevant response time scales and combining a transient
airborne fraction with an equilibrium climate sensitivity.
The TCRE can be compared to the temperature response to emissions
on a time scale of about 1000 years after emissions cease. This can
be estimated from the likely range of equilibrium climate sensitivity
(1.5°C to 4.5°C) and a cumulative CO
2
airborne fraction after about
1109
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
1000 years of about 25 ± 5% (Archer et al., 2009; Joos et al., 2013).
Again combining the extreme values would suggest a range of 0.6°C
to 2.7°C per 1000 PgC, and 1.5°C per 1000 PgC for an ECS of 3°C
and a cumulative airborne fraction of 25%. However, this equilibrium
estimate is based on feedbacks estimated for the present day climate.
Climate and carbon cycle feedbacks may increase substantially on long
time scales and for high cumulative CO
2
emissions (see Section 12.5.3),
introducing large uncertainties in particular on the upper bound. Based
on paleoclimate data and an analytical model, Goodwin et al. (2009)
estimate a long term RF of 1.5 W m
–2
for an emission of 1000 PgC. For
an equilibrium climate sensitivity of 3°C this corresponds to a warming
of 1.2°C on millennial time scales, consistent with the climate carbon
cycle models results discussed above.
The uncertainty in TCRE is caused by the uncertainty in the physical
feedbacks and ocean heat uptake (reflected in TCR) and uncertainties
in carbon cycle feedbacks (affecting the cumulative airborne fraction
of CO
2
). TCRE only characterizes the warming due to CO
2
emissions,
and contributions from non-CO
2
gases need to be considered sepa-
rately when estimating likelihoods to stay below a temperature limit.
Warming as a function of cumulative CO
2
emissions is similar in the
four RCP scenarios, and larger than that due to CO
2
alone, since non-
CO
2
forcings contribute warming in these scenarios (compare Figure
12.45 f, i) (Hajima et al., 2012).
0 500 1000 1500 2000 2500 3000
0
1
2
3
4
5
6
Cumulative CO
2
emissions (PgC)
Transient temperature increase
(°C) (rel. to 1850-1875)
0 500 1000 1500 2000 2500 3000
0
1
2
3
4
5
6
Cumulative CO
2
emissions (PgC)
Transient temperature increase
(°C) (rel. to 1865-1875)
0 500 1000 1500 2000 2500 3000
0
1
2
3
4
5
6
Cumulative CO
2
emissions (PgC)
Transient temperature increase
(°C) (rel. to 1850-1875)
0 500 1000 1500 2000 2500 3000
0
1
2
3
4
5
6
Cumulative CO
2
eq emissions (PgC-eq)
Transient temperature increase
(°C) (rel. to 1850-1875)
Median
66% range
90% range
0
500 1000 1500 2000 2500 3000
0
1
2
3
4
5
6
Cumulative CO
2
emissions to 2200 (PgC)
Peak CO
2
induced warming
(°C) (rel. to per-industrial)
0 500 1000 1500 2000 2500 3000
0
1
2
3
4
5
6
Cumulative diagnosed CO
2
emissions (PgC)
Transient temperature increase
(°C) (rel. to 1850-1875)
0 500 1000 1500 2000 2500 3000
0
1
2
3
4
5
6
Cumulative CO
2
emissions (PgC)
Transient temperature increase
(°C) (rel. to yr 0)
Very likely
Likely
Most likely
1000 1500 2000 2500 3000 3500
0
1
2
3
4
5
6
7
Cumulative Kyoto-gas emissions 2000-2049 (PgCO
2
-eq)
Global mean air surface temperature
(°C) (rel. to 1860-1899)
90%
68%
Medians
Ranges:
b
a
d
g
hi
ef
0 500 1000 1500 2000 2500 3000
0
1
2
3
4
5
6
Cumulative diagnosed CO
2
emissions (PgC)
Transient temperature increase
(°C) (rel. to 1860-1880)
c
FF & industry
(emis-driven)
Total (back-
calculated)
Figure 12.45 | Global temperature change vs. cumulative carbon emissions for different scenarios and models. (a) Transient global temperature increase vs. cumulative CO
2
emis-
sions for Coupled Climate Carbon Cycle Model Intercomparison Project (C
4
MIP) (Matthews et al., 2009). (b) Maximum temperature increase until 2100 vs. cumulative Kyoto-gas
emissions (CO
2
equivalent; note that all other panels are given in C equivalent) (Meinshausen et al., 2009). (c) Transient temperature increase vs. cumulative CO
2
emissions for IPCC
TAR models (red, IPCC TAR Figure 13.3) and IPCC AR4 Earth System Models of Intermediate Complexity (EMICs, black: IPCC AR4 Figure 10.35). (d) As in (a) but for the ENSEMBLES
E1 scenario (Johns et al., 2011). (e) Transient temperature increase for the RCP scenarios based on the Model for the Assessment of Greenhouse Gas-Induced Climate Change
(MAGICC) model constrained to C
4
MIP, observed warming, and the IPCC AR4 climate sensitivity range (Rogelj et al., 2012). (f) Transient temperature change from the CMIP5 1%
yr
–1
concentration driven simulations. (g) Peak CO
2
induced warming vs. cumulative CO
2
emissions to 2200 (Allen et al., 2009; Bowerman et al., 2011). (h) Transient temperature
increase from the new EMIC RCP simulations (Zickfeld et al., 2013). (i) Transient temperature change from the CMIP5 historical and RCP8.5 emission driven simulations (black) and
transient temperature change in all concentration-driven CMIP5 RCP simulations with back-calculated emissions (red). Note that black lines in panel (i) do not include land use CO
2
and that warming in (i) is higher than in (f) due to additional non-CO
2
forcings.
1110
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Box 12.2 | Equilibrium Climate Sensitivity and Transient Climate Response
Equilibrium climate sensitivity (ECS) and transient climate response (TCR) are useful metrics summarizing the global climate system’s
temperature response to an externally imposed radiative forcing (RF). ECS is defined as the equilibrium change in annual mean global
surface temperature following a doubling of the atmospheric CO
2
concentration (see Glossary), while TCR is defined as the annual
mean global surface temperature change at the time of CO
2
doubling following a linear increase in CO
2
forcing over a period of 70 years
(see Glossary). Both metrics have a broader application than these definitions imply: ECS determines the eventual warming in response
to stabilization of atmospheric composition on multi-century time scales, while TCR determines the warming expected at a given time
following any steady increase in forcing over a 50- to 100-year time scale.
ECS and TCR can be estimated from various lines of evidence. The estimates can be based on the values of ECS and TCR diagnosed
from climate models (Section 9.7.1; Table 9.5), or they can be constrained by analysis of feedbacks in climate models (see Section
9.7.2), patterns of mean climate and variability in models compared to observations (Section 9.7.3.3), temperature fluctuations as
reconstructed from paleoclimate archives (Sections 5.3.1 and 5.3.3.2; Box 5.1), observed and modelled short-term perturbations of the
energy balance like those caused by volcanic eruptions (Section 10.8), and the observed surface and ocean temperature trends since
pre-industrial (see Sections 10.8.1 and 10.8.2; Figure 10.20). For many applications, the limitations of the forcing-feedback analysis
framework and the dependence of feedbacks on time scales and the climate state (see Section 12.5.3) must be kept in mind. Some
studies estimate the TCR as the ratio of global mean temperature change to RF (Section 10.8.2.2) (Gregory and Forster, 2008; Padilla
et al., 2011; Schwartz, 2012). Those estimates are scaled by the RF of 2 × CO
2
(3.7 W m
–2
; Myhre et al., 1998) to be comparable to TCR
in the following discussion.
Newer studies of constraints based on the observed warming since
pre-industrial, analysed using simple and intermediate complexity
models, improved statistical methods, and several different and
newer data sets, are assessed in detail in Section 10.8.2. Together
with results from feedback analysis and paleoclimate constraints
(Sections 5.3.1 and 5.3.3.2; Box 5.1), but without considering the
CMIP based evidence, these studies show ECS is likely between
1.5°C to 4.5°C (medium confidence) and extremely unlikely less
than 1.0°C (see Section 10.8.2). A few studies argued for very
low values of climate sensitivity, but many of them have received
criticism in the literature (see Section 10.8.2). Estimates based
on AOGCMs and feedback analysis indicate a range of 2°C to
4.5°C, with the CMIP5 model mean at 3.2°C, similar to CMIP3.
A summary of published ranges and PDFs of ECS is given in Box
12.2, Figure 1. Distributions and ranges for the TCR are shown in
Box 12.2, Figure 2.
Simultaneously imposing different constraints from the observed
warming trends, volcanic eruptions, model climatology, and pale-
oclimate, for example, by using a distribution obtained from the
Last Glacial Maximum as a prior for the 20th century analysis,
yields a more narrow range for climate sensitivity (see Figure
10.20; Section 10.8.2.5) (e.g., Annan and Hargreaves, 2006,
2011b; Hegerl et al., 2006; Aldrin et al., 2012). However, such
methods are sensitive to assumptions of independence of the var-
ious lines of evidence, which might have shared biases (Lemoine,
2010), and the assumption that each individual line of evidence
is unbiased and its uncertainties are captured completely. Expert
elicitations for PDFs of climate sensitivity exist (Morgan and
Keith, 1995; Zickfeld et al., 2010), but have also received some
criticism (Millner et al., 2013). They are not used formally here
because the experts base their opinion on the same studies as we
assess. The peer-reviewed literature provides no consensus on a
Box 12.2, Figure 1 | Probability density functions, distributions and ranges
for equilibrium climate sensitivity, based on Figure 10.20b plus climatological
constraints shown in IPCC AR4 (Meehl et al., 2007b; Box 10.2, Figure 1), and
results from CMIP5 (Table 9.5). The grey shaded range marks the likely 1.5°C to
4.5°C range, and the grey solid line the extremely unlikely less than 1°C, the grey
dashed line the very unlikely greater than 6°C. See Figure 10.20b and Chapter 10
Supplementary Material for full caption and details. Labels refer to studies since
AR4. Full references are given in Section 10.8.
0 1 2 3 4
5
6
7
8 9 10
Equilibrium Climate Sensitivity (°C)
Aldrin et al.(2012)
Bender et al.(2010)
Lewis(2013)
Linetal. (2010)
Lindzen &Choi (2011)
Otto et al.(2013)
Murphy et al.(2009)
Olsonetal. (2012)
Schwartz (2012)
Tomassinietal. (2007)
Sexton et al.(2012)
QUMP
CMIP5
CAM3
MIROC5-CGCM-PPE
CPDN-HadCM3
CMIP3AOGCMs
CMIP5AOGCMs
Chylek & Lohmann (2008)
Hargreaves et al.(2012)
Holden et al.(2010)
K
¨
ohler et al.(2010)
Palaeosens (2012)
Schmittner et al.(2012)
Aldrin et al.(2012)
Libardoni&Forest(2013)
Olsonetal. (2012)
Instrumental
Climatologicalconstraints
Rawmodel range
Palaeoclimate
Combination
(continued on next page)
1111
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Box 12.2 (continued)
formal statistical method to combine different lines of evidence. All methods in general are sensitive to the assumed prior distributions.
These limitations are discussed in detail in Section 10.8.2.
Based on the combined evidence from observed climate change including the observed 20th century warming, climate models, feed-
back analysis and paleoclimate, ECS is likely in the range 1.5°C to 4.5°C with high confidence. The combined evidence increases
the confidence in this final assessment compared to that based
on the observed warming and paleoclimate only. ECS is posi-
tive, extremely unlikely less than 1°C (high confidence), and very
unlikely greater than 6°C (medium confidence). The upper limit of
the likely range is unchanged compared to AR4. The lower limit of
the likely range of 1.5°C is less than the lower limit of 2°C in AR4.
This change reflects the evidence from new studies of observed
temperature change, using the extended records in atmosphere
and ocean. These studies suggest a best fit to the observed sur-
face and ocean warming for ECS values in the lower part of the
likely range. Note that these studies are not purely observation-
al, because they require an estimate of the response to RF from
models. In addition, the uncertainty in ocean heat uptake remains
substantial (see Section 3.2, Box 13.1). Accounting for short
term variability in simple models remains challenging, and it is
important not to give undue weight to any short time period that
might be strongly affected by internal variability (see Box 9.2).
On the other hand, AOGCMs show very good agreement with
observed climatology with ECS values in the upper part of the
1.5°C to 4.5°C range (Section 9.7.3.3), but the simulation of key
feedbacks like clouds remains challenging in those models. The
estimates from the observed warming, paleoclimate, and from
climate models are consistent within their uncertainties, each is
supported by many studies and multiple data sets, and in combi-
nation they provide high confidence for the assessed likely range.
Even though this assessed range is similar to previous reports
(Charney, 1979; IPCC, 2001), confidence today is much higher as
a result of high quality and longer observational records with a
clearer anthropogenic signal, better process understanding, more
and better understood evidence from paleoclimate reconstruc-
tions, and better climate models with higher resolution that cap-
ture many more processes more realistically. Box 12.2 Figure 1
illustrates that all these lines of evidence individually support the
assessed likely range of 1.5°C to 4.5°C.
The tails of the ECS distribution are now better understood. Multiple lines of evidence provide high confidence that an ECS value less
than 1°C is extremely unlikely. The assessment that ECS is very unlikely greater than 6°C is an expert judgment informed by several
lines of evidence. First, the comprehensive climate models used in the CMIP5 exercise produce an ECS range of 2.1°C to 4.7°C (Table
9.5), very similar to CMIP3. Second, comparisons of perturbed-physics ensembles against the observed climate find that models with
ECS values in the range 3°C to 4°C show the smallest errors for many fields (Section 9.7.3.3). Third, there is increasing evidence that the
aerosol RF of the 20th century is not strongly negative, which makes it unlikely that the observed warming was caused by a very large
ECS in response to a very small net forcing. Fourth, multiple and at least partly independent observational constraints from the satellite
period, instrumental period and palaeoclimate studies continue to yield very low probabilities for ECS larger than 6°C, particularly
when including most recent ocean and atmospheric data (see Box 12.2, Figure 1).
Analyses of observations and simulations of the instrumental period are estimating the effective climate sensitivity (a measure of the
strengths of the climate feedbacks today, see Glossary), rather than ECS directly. In some climate models ECS tends to be higher than
the effective climate sensitivity (see Section 12.5.3), because the feedbacks that are represented in the models (water vapour, lapse
Box 12.2, Figure 2 | Probability density functions, distributions and ranges
(5 to 95%) for the transient climate response from different studies, based on
Figure 10.20a, and results from CMIP5 (black histogram; Table 9.5). The grey
shaded range marks the likely 1°C to 2.5°C range, and the grey solid line marks
the extremely unlikely greater than 3°C. See Figure 10.20a and Chapter 10
Supplementary Material for full caption and details. Full references are given
in Section 10.8.
012 3 4 5
0
0.5
1
1.5
Knutti & Tomassini (2008) (b)
Knutti & Tomassini (2008) (a)
Black histogram
CMIP5 models
Dashed lines
AR4 studies
Meinshausen et al (2009)
Harris et al (2013)
Rogelj et al (2012)
Otto et al (2013) (a)
Otto et al (2013) (b)
Tung et al (2008)
Gillett et al (2013)
Stott & Forest (2007)
Gregory & Forster (2008)
Padilla et al (2011)
Libardoni & Forest (2011)
Schwartz (2012)
Transient Climate Response (°C)
Probability / Relative Frequency (°C
−1
)
(continued on next page)
1112
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
12.5.4.3 Conclusions and Limitations
One difficulty with the concepts of climate stabilization and targets is
that stabilization of global temperature does not imply stabilization for
all aspects of the climate system. For example, some models show sig-
nificant hysteresis behaviour in the global water cycle, because global
precipitation depends on both atmospheric CO
2
and temperature (Wu
et al., 2010). Processes related to vegetation changes (Jones et al.,
2009) or changes in the ice sheets (Charbit et al., 2008; Ridley et al.,
2010) as well as ocean acidification, deep ocean warming and asso-
ciated sea level rise (Meehl et al., 2005b; Wigley, 2005; Zickfeld et al.,
2013) (see Figure 12.44d), and potential feedbacks linking, for exam-
ple, ocean and the ice sheets (Gillett et al., 2011; Goelzer et al., 2011),
have their own intrinsic long time scales. Those will result in significant
changes hundreds to thousands of years after global temperature is
stabilized. Thermal expansion, in contrast to global mean temperature,
also depends on the evolution of surface temperature (Stouffer and
Manabe, 1999; Bouttes et al., 2013; Zickfeld et al., 2013).
Box 12.2 (continued)
rate, albedo and clouds) vary with the climate state. On time scales of many centuries, additional feedbacks with their own intrinsic
time scales (e.g., vegetation, ice sheets; see Sections 5.3.3 and 12.5.3) (Jones et al., 2009; Goelzer et al., 2011) may become important
but are not usually modelled. The resulting Earth system sensitivity is less well constrained but likely to be larger than ECS (Hansen et
al., 2008; Rohling et al., 2009; Lunt et al., 2010; Pagani et al., 2010; Rohling and Members, 2012), implying that lower atmospheric CO
2
concentrations are needed to meet a given temperature target on multi-century time scales. A number of caveats, however, apply to
those studies (see Section 12.5.3). Those long-term feedbacks have their own intrinsic time scales, and are less likely to be proportional
to global mean temperature change.
For scenarios of increasing RF, TCR is a more informative indicator of future climate than ECS (Frame et al., 2005; Held et al., 2010). This
assessment concludes with high confidence that the TCR is likely in the range 1°C to 2.5°C, close to the estimated 5 to 95% range of
CMIP5 (1.2°C to 2.4°C; see Table 9.5), is positive and extremely unlikely greater than 3°C. As with the ECS, this is an expert-assessed
range, supported by several different and partly independent lines of evidence, each based on multiple studies, models and data sets.
TCR is estimated from the observed global changes in surface temperature, ocean heat uptake and RF, the detection/attribution studies
identifying the response patterns to increasing GHG concentrations (Section 10.8.1), and the results of CMIP3 and CMIP5 (Section
9.7.1). Estimating TCR suffers from fewer difficulties in terms of state- or time-dependent feedbacks (see Section 12.5.3), and is less
affected by uncertainty as to how much energy is taken up by the ocean. Unlike ECS, the ranges of TCR estimated from the observed
warming and from AOGCMs agree well, increasing our confidence in the assessment of uncertainties in projections over the 21st
century.
Another useful metric relating directly CO
2
emissions to temperature is the transient climate response to cumulative carbon emission
(TCRE) (see Sections 12.5.4 and 10.8.4). This metric is useful to determine the allowed cumulative carbon emissions for stabilization at
a specific global temperature. TCRE is defined as the annual mean global surface temperature change per unit of cumulated CO
2
emis-
sions, usually 1000 PgC, in a scenario with continuing emissions (see Glossary). It considers physical and carbon cycle feedbacks and
uncertainties, but not additional feedbacks associated for example with the release of methane hydrates or large amounts of carbon
from permafrost. The assessment based on climate models as well as the observed warming suggests that the TCRE is likely between
0.8°C to 2.5°C per 1000 PgC (10
12
metric tons of carbon), for cumulative CO
2
emissions less than about 2000 PgC until the time at which
temperatures peak. Under these conditions, and for low to medium estimates of climate sensitivity, the TCRE gives an accurate estimate
of the peak global mean temperature response to cumulated carbon emissions. TCRE has the advantage of directly relating global mean
surface temperature change to CO
2
emissions, but as a result of combining the uncertainty in both TCR and the carbon cycle response,
it is more uncertain. It also ignores non-CO
2
forcings and the fact that other components of the climate system (e.g., sea level rise, ice
sheets) have their own intrinsic time scales, resulting in climate change not avoided by limiting global temperature change.
The simplicity of the concept of a cumulative carbon emission budget
makes it attractive for policy (WBGU, 2009). The principal driver of long
term warming is the total cumulative emission of CO
2
over time. To
limit warming caused by CO
2
emissions to a given temperature target,
cumulative CO
2
emissions from all anthropogenic sources therefore
need to be limited to a certain budget. Higher emissions in earlier dec-
ades simply imply lower emissions by the same amount later on. This
is illustrated in the RCP2.6 scenario in Figure 12.46a/b. Two idealized
emission pathways with initially higher emissions (even sustained at
high level for a decade in one case) eventually lead to the same warm-
ing if emissions are then reduced much more rapidly. Even a stepwise
emission pathway with levels constant at 2010 and zero near mid-cen-
tury would eventually lead to a similar warming as they all have iden-
tical cumulative emissions.
However, several aspects related to the concept of a cumulative carbon
emission budget should be kept in mind. The ratio of global tempera-
ture and cumulative carbon is only approximately constant. It is the
result of an interplay of several compensating carbon cycle and climate
1113
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
feedback processes operating on different time scales (a cancellation of
variations in the increase in RF per ppm of CO
2
, the ocean heat uptake
efficiency and the airborne fraction) (Gregory et al., 2009; Matthews
et al., 2009; Solomon et al., 2009). It depends on the modelled climate
sensitivity and carbon cycle feedbacks. Thus, the allowed emissions for
a given temperature target are uncertain (see Figure 12.45) (Matthews
et al., 2009; Zickfeld et al., 2009; Knutti and Plattner, 2012). Neverthe-
less, the relationship is nearly linear in all models. Most models do not
consider the possibility that long term feedbacks (Hansen et al., 2007;
Knutti and Hegerl, 2008) may be different (see Section 12.5.3). Despite
the fact that stabilization refers to equilibrium, the results assessed
here are primarily relevant for the next few centuries and may differ
for millennial scales. Notably, many of these limitations apply similarly
to other policy targets, for example, stabilizing the atmospheric CO
2
concentration.
Non-CO
2
forcing constituents are important, which requires either
assumptions on how CO
2
emission reductions are linked to changes
in other forcings (Meinshausen et al., 2006; Meinshausen et al., 2009;
McCollum et al., 2013), or separate emission budgets and climate
modelling for short-lived and long-lived gases. So far, many studies
ignored non-CO
2
forcings altogether. Those that consider them find
significant effects, in particular warming of several tenths of a degree
for abrupt reductions in emissions of short-lived species, like aerosols
(Brasseur and Roeckner, 2005; Hare and Meinshausen, 2006; Zickfeld
et al., 2009; Armour and Roe, 2011; Tanaka and Raddatz, 2011) (see
also FAQ 12.3). Other studies, which model reductions that explicitly
target warming from short-lived non-CO
2
species only, find important
short-term cooling benefits shortly after the reduction of these species
(Shindell et al., 2012), but do not extend beyond 2030.
The concept of cumulative carbon also implies that higher initial emis-
sions can be compensated by a faster decline in emissions later or by
negative emissions. However, in the real world short-term and long-
term goals are not independent and mitigation rates are limited by
economic constraints and existing infrastructure (Rive et al., 2007;
Mignone et al., 2008; Meinshausen et al., 2009; Davis et al., 2010;
Friedlingstein et al., 2011; Rogelj et al., 2013). An analysis of 193
published emission pathways with an energy balance model (UNEP,
2010; Rogelj et al., 2011) is shown in Figure 12.46c, d. Those emission
pathways that likely limit warming below 2°C (above pre-industrial)
by 2100 show emissions of about 31 to 46 Pg(CO
2
-eq) yr
–1
and 17 to
23 Pg(CO
2
-eq) yr
–1
by 2020 and 2050, respectively. Median 2010 emis-
sions of all models are 48 Pg(CO
2
-eq) yr
–1
. Note that, as opposed to
Figure 12.46a, b, many scenarios still have positive emissions in 2100.
As these will not be zero immediately after 2100, they imply that the
warming may exceed the target after 2100.
The aspects discussed above do not limit the robustness of the overall
scientific assessment, but highlight factors that need to be considered
when determining cumulative CO
2
emissions consistent with a given
temperature target. In conclusion, taking into account the available
information from multiple lines of evidence (observations, models and
process understanding), the near linear relationship between cumula-
tive CO
2
emissions and peak global mean temperature is well estab-
lished in the literature and robust for cumulative total CO
2
emissions
up to about 2000 PgC. It is consistent with the relationship inferred
from past cumulative CO
2
emissions and observed warming, is sup-
ported by process understanding of the carbon cycle and global energy
balance, and emerges as a robust result from the entire hierarchy of
models.
Using a best estimate for the TCRE would provide a most likely value
for the cumulative CO
2
emissions compatible with stabilization at a
given temperature. However, such a budget would imply about 50%
probability for staying below the temperature target. Higher probabil-
ities for staying below a temperature or concentration target require
significantly lower budgets (Knutti et al., 2005; Meinshausen et al.,
2009; Rogelj et al., 2012). Based on the assessment of TCRE (assum-
ing a normal distribution with a ±1 standard deviation range of 0.8-
2.5°C per 1000 PgC), limiting the warming caused by anthropogenic
CO
2
emissions alone (i.e., ignoring other radiative forcings) to less than
2°C since the period 1861–1880 with a probability of >33%, >50%
and >66%, total CO
2
emissions from all anthropogenic sources would
need to be below a cumulative budget of about 1570 PgC, 1210 PgC
and 1000 PgC since 1870, respectively. An amount of 515 [445 to 585]
PgC was emitted between 1870 and 2011. Accounting for non-CO
2
forcings contributing to peak warming, or requiring a higher likelihood
of temperatures remaining below 2°C, both imply lower cumulative
CO
2
emissions. A possible release of GHGs from permafrost or meth-
ane hydrates, not accounted for in current models, would also further
reduce the anthropogenic CO
2
emissions compatible with a given tem-
perature target. When accounting for the non-CO
2
forcings as in the
RCP scenarios, compatible carbon emissions since 1870 are reduced
to about 900 PgC, 820 PgC and 790 PgC to limit warming to less than
2°C since the period 1861–1880 with a probability of >33%, >50%,
and >66%, respectively. These estimates were derived by computing
the fraction of CMIP5 ESMs and EMICs that stay below 2°C for given
cumulative emissions following RCP8.5, as shown in TFE.8 Figure 1c.
The non-CO
2
forcing in RCP8.5 is higher than in RCP2.6. Because all
likelihood statements in calibrated IPCC language are open intervals,
the provided estimates are thus both conservative and consistent
choices valid for non-CO
2
forcings across all RCP scenarios. There is no
RCP scenario which limits warming to 2°C with probabilities of >33%
or >50%, and which could be used to directly infer compatible cumu-
lative emissions. For a probability of >66% RCP2.6 can be used as a
comparison. Combining the average back-calculated fossil fuel carbon
emissions for RCP2.6 between 2012 and 2100 (270 PgC) with the aver-
age historical estimate of 515 PgC gives a total of 785 PgC, i.e., 790
PgC when rounded to 10 PgC. As the 785 PgC estimate excludes an
explicit assessment of future land-use change emissions, the 790 PgC
value also remains a conservative estimate consistent with the overall
likelihood assessment. The ranges of emissions for these three likeli-
hoods based on the RCP scenarios are rather narrow, as they are based
on a single scenario and on the limited sample of models available
(TFE.8 Figure 1c). In contrast to TCRE they do not include observational
constraints or account for sources of uncertainty not sampled by the
models. The concept of a fixed cumulative CO
2
budget holds not just for
2°C, but for any temperature level explored with models so far (up to
about 5°C; see Figures 12.44 to 12.46), with higher temperature levels
implying larger budgets.
1114
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
T<2°C
2°C<T<3°C
3°C<T<4°C
T>4°C
2000 2010 2020 2030 2040 2050 2060 2070 2080 2090 2100
Likely (>66%) temperature increase (T) during 21st century
from illustrative emission trajectories and RCPs
Total GHG emission levels (GtCO
2
-eq yr
-1
)
2000 2010 2020 2030 2040 2050 2060 2070 2080 2090 2100
Median temperature increase per pathway group (min-max)
and per RCP (median)
Temperature increase rel. to preindustrial (°C)
2000 2010 2020 2030 2040 2050 2060 2070 2080 2090 2100
-2
0
2
4
6
8
10
12
Year
CO
2
emissions from fossil fuel and industry (GtC yr
-1
)
2000 2010 2020 2030 2040 2050 2060 2070 2080 2090 2100
0.8
1
1.2
1.4
1.6
1.8
2
2.2
Year
Year Year
Temperature increase relative to preindustrial (°C)
ab
c
d
-20
0
20
40
60
80
100
120
140
1
0
2
3
4
5
6
12.5.5 Potentially Abrupt or Irreversible Changes
12.5.5.1 Introduction
This report adopts the definition of abrupt climate change used in Syn-
thesis and Assessment Product 3.4 of the U.S. Climate Change Science
Program CCSP (CCSP, 2008b). We define abrupt climate change as a
large-scale change in the climate system that takes place over a few
decades or less, persists (or is anticipated to persist) for at least a few
decades, and causes substantial disruptions in human and natural sys-
tems (see Glossary). Other definitions of abrupt climate change exist.
For example, in the AR4 climate change was defined as abrupt if it
occurred faster than the typical time scale of the responsible forcing.
A number of components or phenomena within the Earth system have
been proposed as potentially possessing critical thresholds (some-
Figure 12.46 | (a) CO
2
emissions for the RCP2.6 scenario (black) and three illustrative modified emission pathways leading to the same warming. (b) Global temperature change
relative to pre-industrial for the pathways shown in panel (a). (c) Grey shaded bands show Integrated Assessment Model (IAM) emission pathways over the 21st century. The
pathways were grouped based on ranges of likely avoided temperature increase in the 21st century. Pathways in the darkest three bands likely stay below 2°C, 3°C, 4°C by 2100,
respectively (see legend), while those in the lightest grey band are higher than that. Emission corridors were defined by, at each year, identifying the 15th to 85th percentile range of
emissions and drawing the corresponding bands across the range. Individual scenarios that follow the upper edge of the bands early on tend to follow the lower edge of the band
later on. Black-white lines show median paths per range. (d) Global temperature relative to pre-industrial for the pathways in (c). (Data in (c) and (d) based on Rogelj et al. (2011).)
Coloured lines in (c) and (d) denote the four RCP scenarios.
times referred to as tipping points (Lenton et al., 2008)), beyond which
abrupt or nonlinear transitions to a different state ensues. The term
irreversibility is used in various ways in the literature. The AR5 report
defines a perturbed state as irreversible on a given time scale if the
recovery time scale from this state due to natural processes is sig-
nificantly longer than the time it takes for the system to reach this
perturbed state (see Glossary). In that context, most aspects of the cli-
mate change resulting from CO
2
emissions are irreversible, due to the
long residence time of the CO
2
perturbation in the atmosphere and the
resulting warming (Solomon et al., 2009). These results are discussed
in Sections 12.5.2 to 12.5.4. Here, we also assess aspects of irreversi-
bility in the context of abrupt change, multiple steady states and hys-
teresis, i.e., the question whether a change (abrupt or not) would be
reversible if the forcing was reversed or removed (e.g., Boucher et al.,
2012). Irreversibility of ice sheets and sea level rise are also assessed
in Chapter 13.
1115
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
In this section we examine the main components or phenomena within
the Earth system that have been proposed in the literature as potential-
ly being susceptible to abrupt or irreversible change (see Table 12.4).
Abrupt changes that arise from nonlinearities within the climate system
are inherently difficult to assess and their timing, if any, of future occur-
rences is difficult to predict. Nevertheless, progress is being made
exploring the potential existence of early warning signs for abrupt cli-
mate change (see e.g., Dakos et al., 2008; Scheffer et al., 2009).
12.5.5.2 The Atlantic Meridional Overturning
EMICs for which the stability has been systematically assessed by
suitably designed hysteresis experiments robustly show a threshold
beyond which the Atlantic thermohaline circulation cannot be sus-
tained (Rahmstorf et al., 2005). This is also the case for one low-reso-
lution ESM (Hawkins et al., 2011). However, proximity to this threshold
is highly model dependent and influenced by factors that are currently
poorly understood. There is some indication that the CMIP3 climate
models may generally overestimate the stability of the Atlantic Ocean
circulation (Hofmann and Rahmstorf, 2009; Drijfhout et al., 2010). In
particular, De Vries and Weber (2005), Dijkstra (2007), Weber et al.
(2007), Huisman et al. (2010), Drijfhout et al. (2010) and Hawkins et
al. (2011) suggest that the sign of net freshwater flux into the Atlantic
transported through its southern boundary via the overturning circu-
lation determines whether or not the AMOC is in a mono-stable or
bi-stable state. For the pre-industrial control climate of most of the
CMIP3 models, Drijfhout et al. (2010) found that the salt flux was nega-
tive (implying a positive freshwater flux), indicating that they were in a
mono-stable regime. However, this is not the case in the CMIP5 models
where Weaver et al. (2012) found that the majority of the models were
in a bi-stable regime during RCP integrations. Observations suggest
that the present day ocean is in a bi-stable regime, thereby allowing
for multiple equilibria and a stable ‘off’ state of the AMOC (Bryden et
al., 2011; Hawkins et al., 2011).
Change in climate
system component
Potentially
abrupt (AR5
definition)
Irreversibility if
forcing reversed
Projected likelihood of 21st century change in scenarios considered
Atlantic MOC collapse Yes Unknown Very unlikely that the AMOC will undergo a rapid transition (high confidence)
Ice sheet collapse No Irreversible for millennia Exceptionally unlikely that either Greenland or West Antarctic Ice sheets
will suffer near-complete disintegration (high confidence)
Permafrost carbon release No Irreversible for millennia Possible that permafrost will become a net source of atmospheric greenhouse gases (low confidence)
Clathrate methane release Yes Irreversible for millennia Very unlikely that methane from clathrates will undergo catastrophic release (high confidence)
Tropical forests dieback Yes Reversible within
centuries
Low confidence in projections of the collapse of large areas of tropical forest
Boreal forests dieback Yes Reversible within
centuries
Low confidence in projections of the collapse of large areas of boreal forest
Disappearance of
summer Arctic sea ice
Yes Reversible within
years to decades
Likely that the Arctic Ocean becomes nearly ice-free in September before mid-cen-
tury under high forcing scenarios such as RCP8.5 (medium confidence)
Long-term droughts Yes Reversible within
years to decades
Low confidence in projections of changes in the frequency and duration of megadroughts
Monsoonal circulation Yes Reversible within
years to decades
Low confidence in projections of a collapse in monsoon circulations
Table 12.4 | Components in the Earth system that have been proposed in the literature as potentially being susceptible to abrupt or irreversible change. Column 2 defines whether
or not a potential change can be considered to be abrupt under the AR5 definition. Column 3 states whether or not the process is irreversible in the context of abrupt change, and
also gives the typical recovery time scales. Column 4 provides an assessment, if possible, of the likelihood of occurrence of abrupt change in the 21st century for the respective
components or phenomena within the Earth system, for the scenarios considered in this chapter.
In addition to the main threshold for a complete breakdown of the
circulation, others may exist that involve more limited changes, such as
a cessation of Labrador Sea deep water formation (Wood et al., 1999).
Rapid melting of the Greenland ice sheet causes increases in freshwa-
ter runoff, potentially weakening the AMOC. None of the CMIP5 sim-
ulations include an interactive ice sheet component. However, Jung-
claus et al. (2006), Mikolajewicz et al. (2007), Driesschaert et al. (2007)
and Hu et al. (2009) found only a slight temporary effect of increased
melt water fluxes on the AMOC, that was either small compared to the
effect of enhanced poleward atmospheric moisture transport or only
noticeable in the most extreme scenarios.
Although many more model simulations have been conducted since
the AR4 under a wide range of forcing scenarios, projections of the
AMOC behaviour have not changed. Based on the available CMIP5
models, EMICs and the literature, it remains very likely that the AMOC
will weaken over the 21st century relative to pre-industrial. Best esti-
mates and ranges for the reduction from CMIP5 are 11% (1 to 24%)
in RCP2.6 and 34% (12 to 54%) in RCP8.5 (Weaver et al., 2012) (see
Section 12.4.7.2, Figure 12.35). But there is low confidence in the mag-
nitude of the weakening. Drijfhout et al. (2012) show that the AMOC
decrease per degree global mean temperature rise varies from 1.5 to
1.9 Sv (10
6
m
3
s
–1
) for the CMIP5 multi-model ensemble members they
considered depending on the scenario, but that the standard deviation
in this regression is almost half the signal.
The FIO-ESM model shows cooling over much of the NH that may be
related to a strong reduction of the AMOC in all RCP scenarios (even
RCP2.6), but the limited output available from the model precludes
an assessment of the response and realism of this response. Hence
it is not included the overall assessment of the likelihood of abrupt
changes.
1116
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
It is unlikely that the AMOC will collapse beyond the end of the 21st
century for the scenarios considered but a collapse beyond the 21st
century for large sustained warming cannot be excluded.There is low
confidence in assessing the evolution of the AMOC beyond the 21st
century. Two of the CMIP5 models revealed an eventual slowdown of
the AMOC to an off state (Figure 12.35). But this did not occur abruptly.
As assessed by Delworth et al. (2008), for an abrupt transition of the
AMOC to occur, the sensitivity of the AMOC to forcing would have
to be far greater that seen in current models. Alternatively, significant
ablation of the Greenland ice sheet greatly exceeding even the most
aggressive of current projections would be required (Swingedouw et
al., 2007; Hu et al., 2009). While neither possibility can be excluded
entirely, it is unlikely that the AMOC will collapse beyond the end of
the 21st century because of global warming based on the models and
range of scenarios considered.
12.5.5.3 Ice Sheets
As detailed in Section 13.4.3, all available modelling studies agree that
the Greenland ice sheet will significantly decrease in area and volume
in a warmer climate as a consequence of increased melt rates not
compensated for by increased snowfall rates and amplified by positive
feedbacks. Conversely, the surface mass balance of the Antarctic ice
sheet is projected to increase in most projections because increased
snowfall rates outweigh melt increase (see Section 13.4.4).
Irreversibility of ice sheet volume and extent changes can arise because
of the surface-elevation feedback that operates when a decrease of the
elevation of the ice sheet induces a decreased surface mass balance
(generally through increased melting), and therefore essentially applies
to Greenland. As detailed in Section 13.4.3.3, several stable states of
the Greenland ice sheet might exist (Charbit et al., 2008; Ridley et al.,
2010; Langen et al., 2012; Robinson et al., 2012; Solgaard and Langen,
2012), and the ice sheet might irreversibly shrink to a stable small-
er state once a warming threshold is crossed for a certain amount of
time, with the critical duration depending on how far the temperature
threshold has been exceeded. Based on the available evidence (see
Section 13.4.3.3), an irreversible decrease of the Greenland ice sheet
due to surface mass balance changes appears very unlikely in the 21st
century but likely on multi-centennial to millennial time scales in the
strongest forcing scenarios.
In theory (Weertman, 1974; Schoof, 2007) ice sheet volume and extent
changes can be abrupt because of the grounding line instability that
can occur in coastal regions where bedrock is retrograde (i.e., sloping
towards the interior of the ice sheet) and below sea level (see Sec-
tion 4.4.4 and Box 13.2). This essentially applies to West Antarctica,
but also to parts of Greenland and East Antarctica. Furthermore, ice
shelf decay induced by oceanic or atmospheric warming might lead to
abruptly accelerated ice flow further inland (De Angelis and Skvarca,
2003). Because ice sheet growth is usually a slow process, such chang-
es could also be irreversible in the definition adopted here. The availa-
ble evidence (see Section 13.4) suggests that it is exceptionally unlikely
that the ice sheets of either Greenland or West Antarctica will suffer a
near-complete disintegration during the 21st century. More generally,
the potential for abrupt and/or irreversible ice sheet changes (or the
initiation thereof) during the 21st century and beyond is discussed in
detail in Sections 13.4.3 and 13.4.4.
12.5.5.4 Permafrost Carbon Storage
Since the IPCC AR4, estimates of the amount of carbon stored in
permafrost have been significantly revised upwards (Tarnocai et al.,
2009), putting the permafrost carbon stock to an equivalent of twice
the atmospheric carbon pool (Dolman et al., 2010). Because of low
carbon input at high latitudes, permafrost carbon is to a large part of
Pleistocene (Zimov et al., 2006) or Holocene (Smith et al., 2004) origin,
and its potential vulnerability is dominated by decomposition (Eglin et
al., 2010). The conjunction of a long carbon accumulation time scale on
one hand and potentially rapid permafrost thawing and carbon decom-
position under warmer climatic conditions (Zimov et al., 2006; Schuur
et al., 2009; Kuhry et al., 2010) on the other hand suggests poten-
tial irreversibility of permafrost carbon decomposition (leading to an
increase of atmospheric CO
2
and/or CH
4
concentrations) on time scales
of hundreds to thousands of years in a warming climate. Indeed, recent
observations (Dorrepaal et al., 2009; Kuhry et al., 2010) suggest that
this process, induced by widespread permafrost warming and thaw-
ing (Romanovsky et al., 2010), might be already occurring. However,
the existing modelling studies of permafrost carbon balance under
future warming that take into account at least some of the essen-
tial permafrost-related processes (Khvorostyanov et al., 2008; Wania
et al., 2009; Koven et al., 2011; Schaefer et al., 2011; MacDougall et
al., 2012; Schneider von Deimling et al., 2012) do not yield coherent
results beyond the fact that present-day permafrost might become a
net emitter of carbon during the 21st century under plausible future
warming scenarios (low confidence). This also reflects an insufficient
understanding of the relevant soil processes during and after perma-
frost thaw, including processes leading to stabilization of unfrozen soil
carbon (Schmidt et al., 2011), and precludes a firm assessment of the
amplitude of irreversible changes in the climate system potentially
related to permafrost degassing and associated global feedbacks at
this stage (see also Sections 6.4.3.4 and 6.4.7.2 and FAQ 6.1).
12.5.5.5 Atmospheric Methane from Terrestrial and Oceanic
Clathrates
Model simulations (Fyke and Weaver, 2006; Reagan and Moridis, 2007;
Lamarque, 2008; Reagan and Moridis, 2009) suggest that clathrate
deposits in shallow regions (in particular at high latitude regions and in
the Gulf of Mexico) are susceptible to destabilization via ocean warm-
ing. However, concomitant sea level rise due to changes in ocean mass
enhances clathrate stability in the ocean (Fyke and Weaver, 2006). A
recent assessment of the potential for a future abrupt release of meth-
ane was undertaken by the U.S. Climate Change Science Program (Syn-
thesis and Assessment Product 3.4 see Brooke et al., 2008). They con-
cluded that it was very unlikely that such a catastrophic release would
occur this century. However, they argued that anthropogenic warming
will very likely lead to enhanced methane emissions from both terres-
trial and oceanic clathrates (Brooke et al., 2008). Although difficult to
formally assess, initial estimates of the 21st century positive feedback
from methane clathrate destabilization are small but not insignificant
(Fyke and Weaver, 2006; Archer, 2007; Lamarque, 2008). Nevertheless,
on multi-millennial time scales, the positive feedback to anthropogenic
1117
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
warming of such methane emissions is potentially larger (Archer and
Buffett, 2005; Archer, 2007; Brooke et al., 2008). Once more, due to the
difference between release and accumulation time scales, such emis-
sions are irreversible. See also FAQ 6.1.
12.5.5.6 Tropical and Boreal Forests
12.5.5.6.1 Tropical forests
In today’s climate, the strongest growth in the Amazon rainforest
occurs during the dry season when strong insolation is combined with
water drawn from underground aquifers that store the previous wet
season’s rainfall (Huete et al., 2006). AOGCMs do not agree about how
the dry season length in the Amazon may change in the future under
the SRES A1 scenario (Bombardi and Carvalho, 2009), but simulations
with coupled regional climate/potential vegetation models are consist-
ent in simulating an increase in dry season length, a 70% reduction in
the areal extent of the rainforest by the end of the 21st century using
the SRES A2 scenario, and an eastward expansion of the Caatinga
vegetation (Cook and Vizy, 2008; Sorensson et al., 2010). In addition,
some models have demonstrated the existence of multiple equilibria of
the tropical South American climate–vegetation system (e.g., Oyama
and Nobre, 2003). The transition could be abrupt when the dry season
becomes too long for the vegetation to survive, although the resilience
of the vegetation to a longer dry period may be increased by the CO
2
fertilization effect (Zelazowski et al., 2011). Deforestation may also
increase dry season length (Costa and Pires, 2010) and drier conditions
increase the likelihood of wildfires that, combined with fire ignition
associated with human activity, can undermine the forest’s resiliency
to climate change (see also Section 6.4.8.1). If climate change brings
drier conditions closer to those supportive of seasonal forests rather
than rainforest, fire can act as a trigger to abruptly and irreversibly
change the ecosystem (Malhi et al., 2009). However, the existence of
refugia is an important determinant of the potential for the re-emer-
gence of the vegetation (Walker et al., 2009).
Analysis of projected change in the climate–biome space of current
vegetation distributions suggest that the risk of Amazonian forest die-
back is small (Malhi et al., 2009), a finding supported by modelling
when strong carbon dioxide fertilization effects on Amazonian vegeta-
tion are assumed (Rammig et al., 2010). However, the strength of CO
2
fertilization on tropical vegetation is poorly known (see Box 6.3). Uncer-
tainty concerning the existence of critical thresholds in the Amazonian
and other tropical rainforests purely driven by climate change therefore
remains high, and so the possibility of a critical threshold being crossed
in precipitation volume cannot be ruled out (Nobre and Borma, 2009;
Good et al., 2011b, 2011c). Nevertheless, there is still some question as
to whether a transition of the Amazonian or other tropical rainforests
into a lower biomass state could result from the combined effects of
limits to carbon fertilization, climate warming, potential precipitation
decline in interaction with the effects of human land use.
12.5.5.6.2 Boreal forest
Evidence from field observations and biogeochemical modelling make it
scientifically conceivable that regions of the boreal forest could tip into
a different vegetation state under climate warming, but uncertainties
on the likelihood of this occurring are very high (Lenton et al., 2008;
Allen et al., 2010). This is mainly due to large gaps in knowledge con-
cerning relevant ecosystemic and plant physiological responses to
warming (Niinemets, 2010). The main response is a potential advance-
ment of the boreal forest northward and the potential transition from
a forest to a woodland or grassland state on its dry southern edges in
the continental interiors, leading to an overall increase in herbaceous
vegetation cover in the affected parts of the boreal zone (Lucht et al.,
2006). The proposed potential mechanisms for decreased forest growth
and/or increased forest mortality are: increased drought stress under
warmer summer conditions in regions with low soil moisture (Barber et
al., 2000; Dulamsuren et al., 2009, 2010); desiccation of saplings with
shallow roots due to summer drought periods in the top soil layers,
causing suppression of forest reproduction (Hogg and Schwarz, 1997);
leaf tissue damage due to high leaf temperatures during peak summer
temperatures under strong climate warming; and increased insect, her-
bivory and subsequent fire damage in damaged or struggling stands
(Dulamsuren et al., 2008). The balance of effects controlling standing
biomass, fire type and frequency, permafrost thaw depth, snow volume
and soil moisture remains uncertain. Although the existence of, and the
thresholds controlling, a potential critical threshold in the boreal forest
are extremely uncertain, its existence cannot at present be ruled out.
12.5.5.7 Sea Ice
Several studies based on observational data or model hindcasts sug-
gest that the rapidly declining summer Arctic sea ice cover might reach
or might already have passed a tipping point (Lindsay and Zhang, 2005;
Wadhams, 2012; Livina and Lenton, 2013). Identifying Arctic sea ice tip-
ping points from the short observational record is difficult due to high
interannual and decadal variability. In some climate projections, the
decrease in summer Arctic sea ice areal coverage is not gradual but is
instead punctuated by 5- to10- year periods of strong ice loss (Holland
et al., 2006; Vavrus et al., 2012; Döscher and Koenigk, 2013). Still, these
abrupt reductions do not necessarily require the existence of a tipping
point in the system or further imply an irreversible behaviour (Amstrup
et al., 2010; Lenton, 2012). The 5- to 10-year events discussed by Hol-
land et al. (2006) arise when large natural climate variability in the
Arctic reinforces the anthropogenically-forced change (Holland et al.,
2008). Positive trends on the same time scale also occur when internal
variability counteracts the forced change until the middle of the 21st
century (Holland et al., 2008; Kay et al., 2011; Vavrus et al., 2012).
Further work using single-column energy-balance models (Merryfield
et al., 2008; Eisenman and Wettlaufer, 2009; Abbot et al., 2011) yielded
mixed results about the possibility of tipping points and bifurcations
in the transition from perennial to seasonal sea ice cover. Thin ice and
snow covers promote strong longwave radiative loss to space and high
ice growth rates (e.g., Bitz and Roe, 2004; Notz, 2009; Eisenman, 2012).
These stabilizing negative feedbacks can be large enough to overcome
the positive surface–albedo feedback and/or cloud feedback, which act
to amplify the forced sea ice response. In such low-order models, the
emergence of multiple stable states with increased climate forcing is a
parameter-dependent feature (Abbot et al., 2011; Eisenman, 2012). For
example, Eisenman (2012) showed with a single-column energy-bal-
ance model that certain parameter choices that cause thicker ice or
warmer ocean under a given climate forcing make the model more
prone to bifurcations and hence irreversible behaviour.
1118
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
The reversibility of sea ice loss with respect to global or hemispher-
ic mean surface temperature change has been directly assessed in
AOGCMs/ESMs by first raising the CO
2
concentration until virtually all
sea ice disappears year-round and then lowering the CO
2
level at the
same rate as during the ramp-up phase until it reaches again the initial
value (Armour et al., 2011; Boucher et al., 2012; Ridley et al., 2012; Li et
al., 2013b). None of these studies show evidence of a bifurcation lead-
ing to irreversible changes in Arctic sea ice. AOGCMs have also been
used to test summer Arctic sea ice recovery after either sudden or very
rapid artificial removal, and all had sea ice return within a few years
(Schröder and Connolley, 2007; Sedláček et al., 2011; Tietsche et al.,
2011). In the Antarctic, as a result of the strong coupling between the
Southern Ocean’s surface and the deep ocean, the sea ice areal cover-
age in some of the models integrated with ramp-up and ramp-down
atmospheric CO
2
concentration exhibits a significant lag relative to the
global or hemispheric mean surface temperature (Ridley et al., 2012;
Li et al., 2013b), so that its changes may be considered irreversible on
centennial time scales.
Diagnostic analyses of a few global climate models have shown abrupt
sea ice losses in the transition from seasonal to year-round Arctic ice-
free conditions after raising CO
2
to very high levels (Winton, 2006b;
Ridley et al., 2008; Li et al., 2013b), but without evidence for irreversi-
ble changes. Winton (2006b, 2008) hypothesized that the small ice cap
instability (North, 1984) could cause such an abrupt transition. With a
low-order Arctic sea ice model, Eisenman and Wettlaufer (2009) also
found an abrupt change behaviour in the transition from seasonal ice
to year-round ice-free conditions, accompanied by an irreversible bifur-
cation to a new stable, annually ice-free state. They concluded that the
cause is a loss of the stabilizing effect of sea ice growth when the ice
season shrinks in time. The Arctic sea ice may thus experience a sharp
transition to annually ice-free conditions, but the irreversible nature of
this transition seems to depend on the model complexity and structure.
In conclusion, rapid summer Arctic sea ice losses are likely to occur in
the transition to seasonally ice-free conditions. These abrupt changes
might have consequences throughout the climate system as noted by
Vavrus et al. (2011) for cloud cover and Lawrence et al. (2008b) for the
high-latitude ground state. Furthermore, the interannual-to-decadal
variability in the summer Arctic sea ice extent is projected to increase in
response to global warming (Holland et al., 2008; Goosse et al., 2009).
These studies suggest that large anomalies in Arctic sea ice areal cov-
erage, like the ones that occurred in 2007 and 2012, might become
increasingly frequent. However, there is little evidence in global climate
models of a tipping point (or critical threshold) in the transition from
a perennially ice-covered to a seasonally ice-free Arctic Ocean beyond
which further sea ice loss is unstoppable and irreversible.
12.5.5.8 Hydrologic Variability: Long-Term Droughts and
Monsoonal Circulation
12.5.5.8.1 Long-term Droughts
As noted in Section 5.5.5, long-term droughts (often called mega-
droughts, see Glossary) are a recurring feature of Holocene paleocli-
mate records in North America, East and South Asia, Europe, Africa and
India. The transitions into and out of the long-term droughts take many
years. Because the long-term droughts all ended, they are not irrevers-
ible. Nonetheless transitions over years to a decade into a state of
long-term drought would have impacts on human and natural systems.
AR4 climate model projections (Milly et al., 2008) and CMIP5 ensem-
bles (Figure 12.23) both suggest widespread drying and drought across
most of southwestern North America and many other subtropical
regions by the mid to late 21st century (see Section 12.4.5), although
without abrupt change. Some studies suggest that this subtropical
drying may have already begun in southwestern North America (Seager
et al., 2007; Seidel and Randel, 2007; Barnett et al., 2008; Pierce et al.,
2008). More recent studies (Hoerling et al., 2010; Seager and Vecchi,
2010; Dai, 2011; Seager and Naik, 2012) suggest that regional reduc-
tions in precipitation are due primarily to internal variability and that
the anthropogenic forced trends are currently weak in comparison.
While previous long-term droughts in southwest North America arose
from natural causes, climate models project that this region will under-
go progressive aridification as part of a general drying and poleward
expansion of the subtropical dry zones driven by rising GHGs (Held and
Soden, 2006; Seager et al., 2007; Seager and Vecchi, 2010). The models
project the aridification to intensify steadily as RF and global warm-
ing progress without abrupt changes. Because of the very long life-
time of the anthropogenic atmospheric CO
2
perturbation, such drying
induced by global warming would be largely irreversible on millennium
time scale (Solomon et al., 2009; Frölicher and Joos, 2010; Gillett et
al., 2011) (see Sections 12.5.2 and 12.5.4). For example, Solomon et
al. (2009) found in a simulation where atmospheric CO
2
increases to
600 ppm followed by zero emissions, that the 15% reduction in pre-
cipitation in areas such as southwest North America, southern Europe
and western Australia would persist long after emissions ceased. This,
however, is largely a consequence of the warming persisting for centu-
ries after emissions cease rather than an irreversible behaviour of the
water cycle itself.
12.5.5.8.2 Monsoonal circulation
Climate model simulations and paleo-reconstructions provide evidence
of past abrupt changes in Saharan vegetation, with the ‘green Sahara’
conditions (Hoelzmann et al., 1998) of the African Humid Period (AHP)
during the mid-Holocene serving as the most recent example. However,
Mitchell (1990) and Claussen et al. (2003) note that the mid-Holocene is
not a direct analogue for future GHG-induced climate change since the
forcings are different: a increased shortwave forcing in the NH summer
versus a globally and seasonally uniform atmospheric CO
2
increase,
respectively. Paleoclimate examples suggest that a strong radiative
or SST forcing is needed to achieve a rapid climate change, and that
the rapid changes are reversible when the forcing is withdrawn. Both
the abrupt onset and termination of the AHP were triggered when
northern African summer insolation was 4.2% higher than present
day, representing a local increase of about 19 W m
–2
(deMenocal et
al., 2000). Note that the globally averaged radiative anthropogenic
forcing from 1750 to 2011 (Table 8.6) is small compared to this local
increase in insolation. A rapid Saharan greening has been simulated in
a climate model of intermediate complexity forced by a rapid increase
in atmospheric CO
2
, with the overall extent of greening depending on
the equilibrium atmospheric CO
2
level reached (Claussen et al., 2003).
1119
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Abrupt Saharan vegetation changes of the Younger Dryas are linked
with a rapid AMOC weakening which is considered very unlikely during
the 21st century and unlikely beyond that as a consequence of global
warming.
Studies with conceptual models (Zickfeld et al., 2005; Levermann et
al., 2009) have shown that the Indian summer monsoon can operate
in two stable regimes: besides the ‘wet’ summer monsoon, a stable
state exists which is characterized by low precipitation over India.
These studies suggest that any perturbation of the radiative budget
that tends to weaken the driving pressure gradient has the potential to
induce abrupt transitions between these two regimes.
Numerous studies with coupled ocean–atmosphere models have
explored the potential impact of anthropogenic forcing on the Indian
monsoon (see also Section 14.2). When forced with anticipated increas-
es in GHG concentrations, the majority of these studies show an inten-
sification of the rainfall associated with the Indian summer monsoon
(Meehl and Washington, 1993; Kitoh et al., 1997; Douville et al., 2000;
Hu et al., 2000; May, 2002; Ueda et al., 2006; Kripalani et al., 2007; Sto-
wasser et al., 2009; Cherchi et al., 2010). Despite the intensification of
precipitation, several of these modelling studies show a weakening of
the summer monsoon circulation (Kitoh et al., 1997; May, 2002; Ueda
et al., 2006; Kripalani et al., 2007; Stowasser et al., 2009; Cherchi et al.,
2010). The net effect is nevertheless an increase of precipitation due to
enhanced moisture transport into the Asian monsoon region (Ueda et
al., 2006). In recent years, studies with GCMs have also explored the
direct effect of aerosol forcing on the Indian monsoon (Lau et al., 2006;
Meehl et al., 2008; Randles and Ramaswamy, 2008; Collier and Zhang,
2009). Considering absorbing aerosols (black carbon) only, Meehl et
al. (2008) found an increase in pre-monsoonal precipitation, but a
decrease in summer monsoon precipitation over parts of South Asia. In
contrast, Lau et al. (2006) found an increase in May–June–July precipi-
tation in that region. If an increase in scattering aerosols only is consid-
ered, the monsoon circulation weakens and precipitation is inhibited
(Randles and Ramaswamy, 2008). More recently, Bollasina et al. (2011)
showed that anthropogenic aerosols played a fundamental role in driv-
ing the recent observed weakening of the summer monsoon. Given
that the effect of increased atmospheric regional loading of aerosols
is opposed by the concomitant increases in GHG concentrations, it is
unlikely that an abrupt transition to the dry summer monsoon regime
will be triggered in the 21st century.
Acknowledgements
We especially acknowledge the input of Contributing Authors Urs
Beyerle for maintaining the database of CMIP5 output, Jan Sedláček
for producing a large number of CMIP5 figures, and Joeri Rogelj for
preparing synthesis figures. Chapter technical assistants Oliver Stebler,
Franziska Gerber and Barbara Aellig, provided great help in assembling
the chapter and Sébastien Denvil and Jérôme Raciazek provided tech-
nical assistance in downloading the CMIP5 data.
1120
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
References
Abbot, D. S., M. Silber, and R. T. Pierrehumbert, 2011: Bifurcations leading to summer
Arctic sea ice loss. J. Geophys. Res., 116, D19120.
Abe, M., H. Shiogama, T. Nozawa, and S. Emori, 2011: Estimation of future surface
temperature changes constrained using the future-present correlated modes in
inter-model variability of CMIP3 multimodel simulations. J. Geophys. Res., 116,
D18104.
Adachi, Y., et al., 2013: Basic performance of a new earth system model of the
Meteorological Research Institute (MRI-ESM1). Papers Meteorol. Geophys.,
doi:10.2467/mripapers.64.
Adams, P. J., J. H. Seinfeld, D. Koch, L. Mickley, and D. Jacob, 2001: General circulation
model assessment of direct radiative forcing by the sulfate-nitrate-ammonium-
water inorganic aerosol system. J. Geophys. Res., 106, 1097–1111.
Adler, R. F., G. J. Gu, J. J. Wang, G. J. Huffman, S. Curtis, and D. Bolvin, 2008:
Relationships between global precipitation and surface temperature on
interannual and longer timescales (1979–2006). J. Geophys. Res., 113, D22104.
Aldrin, M., M. Holden, P. Guttorp, R. B. Skeie, G. Myhre, and T. K. Berntsen, 2012:
Bayesian estimation of climate sensitivity based on a simple climate model
fitted to observations of hemispheric temperatures and global ocean heat
content. Environmetrics, 23, 253–271.
Alexander, L. V., and J. M. Arblaster, 2009: Assessing trends in observed and modelled
climate extremes over Australia in relation to future projections. Int. J. Climatol.,
29, 417–435.
Alexander, L. V., et al., 2006: Global observed changes in daily climate extremes of
temperature and precipitation. J. Geophys. Res., 111, D05109.
Alexeev, V., and C. Jackson, 2012: Polar amplification: Is atmospheric heat transport
important? Clim. Dyn., doi:10.1007/s00382-012-1601-z.
Alexeev, V., D. Nicolsky, V. Romanovsky, and D. Lawrence, 2007: An evaluation of
deep soil configurations in the CLM3 for improved representation of permafrost.
Geophys. Res. Lett., 34, L09502.
Alexeev, V. A., P. L. Langen, and J. R. Bates, 2005: Polar amplification of surface
warming on an aquaplanet in “ghost forcing” experiments without sea ice
feedbacks. Clim. Dyn., 24, 655–666.
Allan, R., and B. Soden, 2008: Atmospheric warming and the amplification of
precipitation extremes. Science, 321, 1481–1484.
Allan, R. P., 2012: Regime dependent changes in global precipitation. Clim. Dyn.,
doi:10.1007/s00382-011-1134-x.
Allen, C., et al., 2010: A global overview of drought and heat-induced tree mortality
reveals emerging climate change risks for forests. Forest Ecol. Manage., 259,
660–684.
Allen, M. R., and W. J. Ingram, 2002: Constraints on future changes in climate and
the hydrologic cycle. Nature, 419, 224–232.
Allen, M. R., D. J. Frame, C. Huntingford, C. D. Jones, J. A. Lowe, M. Meinshausen,
and N. Meinshausen, 2009: Warming caused by cumulative carbon emissions
towards the trillionth tonne. Nature, 458, 1163–1166.
Allen, R. J., and S. C. Sherwood, 2008: Warming maximum in the tropical upper
troposphere deduced from thermal winds. Nature Geosci., 1, 399–403.
Allen, R. J., and S. C. Sherwood, 2010: Aerosol-cloud semi-direct effect and land-sea
temperature contrast in a GCM. Geophys. Res. Lett., 37, L07702.
Allen, R. J., S. C. Sherwood, J. R. Norris, and C. S. Zender, 2012: Recent Northern
Hemisphere tropical expansion primarily driven by black carbon and tropospheric
ozone. Nature, 485, 350–354.
Amstrup, S., E. DeWeaver, D. Douglas, B. Marcot, G. Durner, C. Bitz, and D. Bailey,
2010: Greenhouse gas mitigation can reduce sea-ice loss and increase polar
bear persistence. Nature, 468, 955–958.
Andrews, T., and P. M. Forster, 2008: CO
2
forcing induces semi-direct effects with
consequences for climate feedback interpretations. Geophys. Res. Lett., 35,
L04802.
Andrews, T., P. M. Forster, and J. M. Gregory, 2009: A surface energy perspective on
climate change. J. Clim., 22, 2557–2570.
Andrews, T., P. Forster, O. Boucher, N. Bellouin, and A. Jones, 2010: Precipitation,
radiative forcing and global temperature change. Geophys. Res. Lett., 37,
L14701.
Annan, J. D., and J. C. Hargreaves, 2006: Using multiple observationally-based
constraints to estimate climate sensitivity. Geophys. Res. Lett., 33, L06704.
Annan, J. D., and J. C. Hargreaves, 2010: Reliability of the CMIP3 ensemble. Geophys.
Res. Lett., 37, L02703.
Annan, J. D., and J. C. Hargreaves, 2011a: Understanding the CMIP3 multi-model
ensemble. J. Clim., 24, 4529–4538.
Annan, J. D., and J. C. Hargreaves, 2011b: On the generation and interpretation of
probabilistic estimates of climate sensitivity. Clim. Change, 104, 423–436.
Arblaster, J. M., G. A. Meehl, and D. J. Karoly, 2011: Future climate change in the
Southern Hemisphere: Competing effects of ozone and greenhouse gases.
Geophys. Res. Lett., 38, L02701.
Archer, D., 2007: Methane hydrate stability and anthropogenic climate change.
Biogeosciences, 4, 521–544.
Archer, D., and B. Buffett, 2005: Time-dependent response of the global ocean
clathrate reservoir to climatic and anthropogenic forcing. Geochem. Geophys.
Geosyst., 6, Q03002.
Archer, D., et al., 2009: Atmospheric lifetime of fossil fuel carbon dioxide. Annu. Rev.
Earth Planet. Sci., 37, 117–134.
Armour, K., and G. Roe, 2011: Climate commitment in an uncertain world. Geophys.
Res. Lett., 38, L01707.
Armour, K., I. Eisenman, E. Blanchard-Wrigglesworth, K. McCusker, and C. Bitz, 2011:
The reversibility of sea ice loss in a state-of-the-art climate model. Geophys. Res.
Lett., 38, L16705.
Arora, V. K., et al., 2011: Carbon emission limits required to satisfy future
representative concentration pathways of greenhouse gases. Geophys. Res.
Lett., 38, L05805.
Arzel, O., T. Fichefet, and H. Goosse, 2006: Sea ice evolution over the 20th and 21st
centuries as simulated by current AOGCMs. Ocean Model., 12, 401–415.
Augustsson, T., and V. Ramanathan, 1977: Radiative-convective model study of CO
2
climate problem. J. Atmos. Sci., 34, 448–451.
Bala, G., K. Caldeira, and R. Nemani, 2010: Fast versus slow response in climate
change: Implications for the global hydrological cycle. Clim. Dyn., 35, 423–434.
Baldwin, M. P., M. Dameris, and T. G. Shepherd, 2007: Atmosphere—How will the
stratosphere affect climate change? Science, 316, 1576–1577.
Ballester, J., F. Giorgi, and X. Rodo, 2010a: Changes in European temperature
extremes can be predicted from changes in PDF central statistics. Clim. Change,
98, 277–284.
Ballester, J., X. Rodo, and F. Giorgi, 2010b: Future changes in Central Europe heat
waves expected to mostly follow summer mean warming. Clim. Dyn., 35, 1191–
1205.
Banks, H. T., and J. M. Gregory, 2006: Mechanisms of ocean heat uptake in a coupled
climate model and the implications for tracer based predictions of ocean heat
uptake. Geophys. Res. Lett., 33, L07608.
Bao, Q., et al., 2013: The Flexible Global Ocean-Atmosphere-Land system model,
Spectral Version 2: FGOALS-s2. Adv. Atmos. Sci., 30, 561–576.
Barber, V., G. Juday, and B. Finney, 2000: Reduced growth of Alaskan white spruce
in the twentieth century from temperature-induced drought stress. Nature, 405,
668–673.
Barnes, E. A., and L. M. Polvani, 2013: Response of the midlatitude jets and of
their variability to increased greenhouse gases in the CMIP5 models. J. Clim.,
doi:10.1175/JCLI-D-12-00536.1.
Barnett, D. N., S. J. Brown, J. M. Murphy, D. M. H. Sexton, and M. J. Webb, 2006:
Quantifying uncertainty in changes in extreme event frequency in response
to doubled CO
2
using a large ensemble of GCM simulations. Clim. Dyn., 26,
489–511.
Barnett, T., and D. Pierce, 2008: When will Lake Mead go dry? Water Resour. Res.,
44, W03201.
Barnett, T. P., et al., 2008: Human-induced changes in the hydrology of the western
United States. Science, 319, 1080–1083.
Barriopedro, D., E. M. Fischer, J. Luterbacher, R. Trigo, and R. Garcia-Herrera, 2011:
The hot summer of 2010: Redrawing the temperature record map of Europe.
Science, 332, 220–224.
Bekryaev, R. V., I. V. Polyakov, and V. A. Alexeev, 2010: Role of polar amplification
in long-term surface air temperature variations and modern Arctic warming. J.
Clim., 23, 3888–3906.
Bellouin, N., J. Rae, A. Jones, C. Johnson, J. Haywood, and O. Boucher, 2011: Aerosol
forcing in the Hadley Centre CMIP5 simulations and the role of ammonium
nitrate. J. Geophys. Res., 116, D20206.
Bengtsson, L., K. I. Hodges, and E. Roeckner, 2006: Storm tracks and climate change.
J. Clim., 19, 3518–3543.
1121
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Bengtsson, L., K. I. Hodges, and N. Keenlyside, 2009: Will extratropical storms
intensify in a warmer climate? J. Clim., 22, 2276–2301.
Berg, P., J. O. Haerter, P. Thejll, C. Piani, S. Hagemann, and J. H. Christensen, 2009:
Seasonal characteristics of the relationship between daily precipitation intensity
and surface temperature. J. Geophys. Res., 114, D18102.
Betts, R., et al., 2007: Projected increase in continental runoff due to plant responses
to increasing carbon dioxide. Nature, 448, 1037–1041.
Bitz, C., and G. Roe, 2004: A mechanism for the high rate of sea ice thinning in the
Arctic Ocean. J. Clim., 17, 3623–3632.
Bitz, C., and Q. Fu, 2008: Arctic warming aloft is data set dependent. Nature, 455,
E3–E4.
Bitz, C. M., 2008: Some aspects of uncertainty in predicting sea ice thinning. In:
Arctic Sea Ice Decline: Observations, Projections, Mechanisms, and Implications
[E. T. DeWeaver, C. M. Bitz and L. B. Tremblay (eds.)]. American Geophysical
Union, Washington, DC, pp. 63–76.
Bitz, C. M., J. K. Ridley, M. M. Holland, and H. Cattle, 2012: Global climate models
and 20th and 21st century Arctic climate change. In: Arctic Climate Change – The
ACSYS Decade and Beyond [P. Lemke (ed.)]. Springer Science+Business Media,
Dordrecht, Netherlands, pp. 405–436.
Boberg, F., P. Berg, P. Thejll, W. Gutowski, and J. Christensen, 2010: Improved
confidence in climate change projections of precipitation evaluated using daily
statistics from the PRUDENCE ensemble. Clim. Dyn., 35, 1097–1106.
Boé, J., and L. Terray, 2008: Uncertainties in summer evapotranspiration changes
over Europe and implications for regional climate change. Geophys. Res. Lett.,
35, L05702.
Boé, J., A. Hall, and X. Qu, 2009a: Current GCMs’ unrealistic negative feedback in the
Arctic. J. Clim., 22, 4682–4695.
Boé, J. L., A. Hall, and X. Qu, 2009b: September sea-ice cover in the Arctic Ocean
projected to vanish by 2100. Nature Geosci., 2, 341–343.
Boer, G. J., 1993: Climate change and the regulation of the surface moisture and
energy budgets. Clim. Dyn., 8, 225–239.
Boer, G. J., 2011: The ratio of land to ocean temperature change under global
warming. Clim. Dyn., 37, 2253–2270.
Boer, G. J., and B. Yu, 2003: Climate sensitivity and response. Clim. Dyn., 20, 415–429.
Boer, G. J., K. Hamilton, and W. Zhu, 2005: Climate sensitivity and climate change
under strong forcing. Clim. Dyn., 24, 685–700.
Bollasina, M. A., Y. Ming, and V. Ramaswamy, 2011: Anthropogenic aerosols and the
weakening of the South Asian summer monsoon. Science, 334, 502–505.
Bombardi, R., and L. Carvalho, 2009: IPCC global coupled model simulations of the
South America monsoon system. Clim. Dyn., 33, 893–916.
ning, C., A. Dispert, M. Visbeck, S. Rintoul, and F. Schwarzkopf, 2008: The response
of the Antarctic Circumpolar Current to recent climate change. Nature Geosci.,
1, 864–869.
Bony, S., and J. L. Dufresne, 2005: Marine boundary layer clouds at the heart of
tropical cloud feedback uncertainties in climate models. Geophys. Res. Lett., 32,
L20806.
Bony, S., G. Bellon, D. Klocke, S. Sherwood, S. Fermepin, and S. Denvil, 2013: Robust
direct effect of carbon dioxide on tropical circulation and regional precipitation.
Nature Geosci., doi:10.1038/ngeo1799.
Bony, S., et al., 2006: How well do we understand and evaluate climate change
feedback processes? J. Clim., 19, 3445–3482.
Booth, B. B. B., et al., 2012: High sensitivity of future global warming to land carbon
cycle processes. Environ. Res. Lett., 7, 024002.
Boucher, O., et al., 2012: ReversibilityinanEarthSystemmodelinresponsetoCO
2
concentrationchanges. Environ. Res. Lett., 7, 024013.
Bouttes, N., J. M. Gregory, and J. A. Lowe, 2013: The reversibility of sea level rise. J.
Clim., 26, 2502–2513.
Bowerman, N., D. Frame, C. Huntingford, J. Lowe, and M. Allen, 2011: Cumulative
carbon emissions, emissions floors and short-term rates of warming: Implications
for policy. Philos. Trans. R. Soc. A, 369, 45–66.
Bracegirdle, T., and D. Stephenson, 2012: Higher precision estimates of regional polar
warming by ensemble regression of climate model projections. Clim. Dyn., 39,
2805–2821.
Bracegirdle, T., W. Connolley, and J. Turner, 2008: Antarctic climate change over the
twenty first century. J. Geophys. Res., 113, D03103.
Bracegirdle, T. J., et al., 2013: Assessment of surface winds over the Atlantic, Indian,
and Pacific Ocean sectors of the Southern Ocean in CMIP5 models: Historical
bias, forcing response, and state dependence. J. Geophys. Res., 118, 547–562.
Brasseur, G., and E. Roeckner, 2005: Impact of improved air quality on the future
evolution of climate. Geophys. Res. Lett., 32, L23704.
Brient, F., and S. Bony, 2013: Interpretation of the positive low-cloud feedback
predicted by a climate model under global warming. Clim. Dyn., 40, 2415–2431.
Brierley, C. M., M. Collins, and A. J. Thorpe, 2010: The impact of perturbations to
ocean-model parameters on climate and climate change in a coupled model.
Clim. Dyn., 34, 325–343.
Bromwich, D. H., J. P. Nicolas, A. J. Monaghan, M. A. Lazzara, L. M. Keller, G. A.
Weidner, and A. B. Wilson, 2013: Central West Antarctica among the most rapidly
warming regions on Earth. Nature Geosci., 6, 139–145.
Brooke, E., D. Archer, E. Dlugokencky, S. Frolking, and D. Lawrence, 2008: Potential
for abrupt changes in atmospheric methane. Abrupt Climate Change: A Report
by the U.S. Climate Change Science Program and the Subcommittee on Global
Change Research. U.S. Geological Survey, Washington, DC, pp. 163–201.
Brooks, H. E., 2009: Proximity soundings for severe convection for Europe and the
United States from reanalysis data. Atmos. Res., 93, 546–553.
Brooks, H. E., 2013: Severe thunderstorms and climate change. Atmos. Res., 123,
129–138.
Brooks, H. E., J. W. Lee, and J. P. Craven, 2003: The spatial distribution of severe
thunderstorm and tornado environments from global reanalysis data. Atmos.
Res., 67–68, 73–94.
Brovkin, V., et al., 2013: Effect of anthropogenic land-use and land cover changes
on climate and land carbon storage in CMIP5 projections for the 21st century. J.
Clim., doi:10.1175/JCLI-D-12–00623.1.
Brown, R., and P. Mote, 2009: The response of Northern Hemisphere snow cover to
a changing climate. J. Clim., 22, 2124–2145.
Brown, R. D., and D. A. Robinson, 2011: Northern Hemisphere spring snow cover
variability and change over 1922–2010 including an assessment of uncertainty.
Cryosphere, 5, 219–229.
Brutel-Vuilmet, C., M. Menegoz, and G. Krinner, 2013: An analysis of present and
future seasonal Northern Hemisphere land snow cover simulated by CMIP5
coupled climate models. Cryosphere, 7, 67–80.
Bryan, K., F. G. Komro, S. Manabe, and M. J. Spelman, 1982: Transient climate
response to increasing atmospheric carbon-dioxide. Science, 215, 56–58.
Bryden, H. L., B. A. King, and G. D. McCarthy, 2011: South Atlantic overturning
circulation at 24S. J. Mar. Res., 69, 38–55.
Burke, E., and S. Brown, 2008: Evaluating uncertainties in the projection of future
drought. J. Hydrometeorol., 9, 292–299.
Burke, E. J., C. D. Jones, and C. D. Koven, 2012: Estimating the permafrost-carbon-
climate response in the CMIP5 climate models using a simplified approach. J.
Clim., doi:10.1175/JCLI-D-12-00550.1.
Buser, C. M., H. R. Kunsch, D. Luthi, M. Wild, and C. Schär, 2009: Bayesian multi-
model projection of climate: Bias assumptions and interannual variability. Clim.
Dyn., 33, 849–868.
Butchart, N., and A. A. Scaife, 2001: Removal of chlorofluorocarbons by increased
mass exchange between the stratosphere and troposphere in a changing
climate. Nature, 410, 799–802.
Butchart, N., et al., 2006: Simulations of anthropogenic change in the strength of the
Brewer-Dobson circulation. Clim. Dyn., 27, 727–741.
Butchart, N., et al., 2010: Chemistry-climate model simulations of twenty-first
century stratospheric climate and circulation changes. J. Clim., 23, 5349–5374.
Butler, A. H., D. W. J. Thompson, and R. Heikes, 2010: The steady-state atmospheric
circulation response to climate change-like thermal forcings in a simple General
Circulation Model. J. Clim., 23, 3474–3496.
Cabre, M. F., S. A. Solman, and M. N. Nunez, 2010: Creating regional climate change
scenarios over southern South America for the 2020’s and 2050’s using the
pattern scaling technique: Validity and limitations. Clim. Change, 98, 449–469.
Caesar, J., and J. A. Lowe, 2012: Comparing the impacts of mitigation versus non-
intervention scenarios on future temperature and precipitation extremes in the
HadGEM2 climate model. J. Geophys. Res., 117, D15109.
Cagnazzo, C., E. Manzini, P. G. Fogli, M. Vichi, and P. Davini, 2013: Role of stratospheric
dynamics in the ozone–carbon connection in the Southern Hemisphere. Clim.
Dyn., doi:10.1007/s00382-013-1745-5.
Cai, M., 2005: Dynamical amplification of polar warming. Geophys. Res. Lett., 32,
L22710.
Caldeira, K., and J. F. Kasting, 1993: Insensitivity of global warming potentials to
carbon-dioxide emission scenarios. Nature, 366, 251–253.
Caldwell, P., and C. S. Bretherton, 2009: Response of a subtropical stratocumulus-
capped mixed layer to climate and aerosol changes. J. Clim., 22, 20–38.
1122
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Calvo, N., and R. R. Garcia, 2009: Wave forcing of the tropical upwelling in the lower
stratosphere under increasing concentrations of greenhouse gases. J. Atmos.
Sci., 66, 3184–3196.
Calvo, N., R. R. Garcia, D. R. Marsh, M. J. Mills, D. E. Kinnison, and P. J. Young,
2012: Reconciling modeled and observed temperature trends over Antarctica.
Geophys. Res. Lett., 39, L16803.
Cao, L., and K. Caldeira, 2010: Atmospheric carbon dioxide removal: Long-term
consequences and commitment. Environ. Res. Lett., 5, 024011.
Cao, L., G. Bala, and K. Caldeira, 2012: Climate response to changes in atmospheric
carbon dioxide and solar irradiance on the time scale of days to weeks. Environ.
Res. Lett., 7, 034015.
Capotondi, A., M. Alexander, N. Bond, E. Curchitser, and J. Scott, 2012: Enhanced
upper ocean stratification with climate change in the CMIP3 models. J. Geophys.
Res., 117, C04031.
Cariolle, D., and H. Teyssedre, 2007: A revised linear ozone photochemistry
parameterization for use in transport and general circulation models: Multi-
annual simulations. Atmos. Chem. Phys., 7, 2183–2196.
Carslaw, K., O. Boucher, D. Spracklen, G. Mann, J. Rae, S. Woodward, and M. Kulmala,
2010: A review of natural aerosol interactions and feedbacks within the Earth
system. Atmos. Chem. Phys., 10, 1701–1737.
Catto, J. L., L. C. Shaffrey, and K. I. Hodges, 2011: Northern Hemisphere extratropical
cyclones in a warming climate in the HiGEM high-resolution climate model. J.
Clim., 24, 5336–5352.
CCSP, 2008a: Weather and Climate Extremes in a Changing Climate: A Report by the
U.S. Climate Change Science Program and the Subcommittee on Global Change
Research. Department of Commerce, NOAAs National Climatic Data Center,
College Park, MD, 164 pp.
CCSP, 2008b: Abrupt Climate Change. A Report by the U.S. Climate Change Science
Program and the Subcommittee on Global Change Research. U.S. Geological
Survey, Washington, DC, 459 pp.
Cess, R., et al., 1990: Intercomparison and interpretation of climate feedback
processes in 19 atmospheric general-circulation models. J. Geophys. Res., 95,
16601–16615.
Chadwick, R., I. Boutle, and G. Martin, 2012: Spatial patterns of precipitation change
in CMIP5: Why the rich don’t get richer in the Tropics. J. Clim., doi:10.1175/JCLI-
D-12-00543.1.
Chadwick, R., P. Wu, P. Good, and T. Andrews, 2013: Asymmetries in tropical rainfall
and circulation patterns in idealised CO
2
removal experiments. Clim. Dyn., 40,
295–316.
Chang, E. K. M., Y. Guo, and X. Xia, 2012a: CMIP5 multimodel ensemble projection
of storm track change under global warming. J. Geophys. Res., 117, D23118.
Chang, E. K. M., Y. Guo, X. Xia, and M. Zheng, 2012b: Storm track activity in IPCC
AR4/CMIP3 model simulations. J. Clim., 26, 246–260.
Chapin, F., et al., 2005: Role of land-surface changes in Arctic summer warming.
Science, 310, 657–660.
Charbit, S., D. Paillard, and G. Ramstein, 2008: Amount of CO
2
emissions irreversibly
leading to the total melting of Greenland. Geophys. Res. Lett., 35, L12503.
Charney, J. G., 1979: Carbon Dioxide and Climate: A Scientific Assessment. National
Academies of Science Press, Washington, DC, 22 pp.
Chen, C. T., and T. Knutson, 2008: On the verification and comparison of extreme
rainfall indices from climate models. J. Clim., 21, 1605–1621.
Chen, G., J. Lu, and D. M. W. Frierson, 2008: Phase speed spectra and the latitude
of surface westerlies: Interannual variability and global warming trend. J. Clim.,
21, 5942–5959.
Cherchi, A., A. Alessandri, S. Masina, and A. Navarra, 2010: Effect of increasing CO
2
levels on monsoons. Clim. Dyn., 37, 83–101.
Choi, D. H., J. S. Kug, W. T. Kwon, F. F. Jin, H. J. Baek, and S. K. Min, 2010: Arctic
Oscillation responses to greenhouse warming and role of synoptic eddy
feedback. J. Geophys. Res. Atmos., 115, D17103.
Chou, C., and J. D. Neelin, 2004: Mechanisms of global warming impacts on regional
tropical precipitation. J. Clim., 17, 2688–2701.
Chou, C., and C. Chen, 2010: Depth of convection and the weakening of tropical
circulation in global warming. J. Clim., 23, 3019–3030.
Chou, C., and C.-W. Lan, 2012: Changes in the annual range of precipitation under
global warming. J. Clim., 25, 222–235.
Chou, C., J. D. Neelin, J. Y. Tu, and C. T. Chen, 2006: Regional tropical precipitation
change mechanisms in ECHAM4/OPYC3 under global warming. J. Clim., 19,
4207–4223.
Chou, C., J. D. Neelin, C. A. Chen, and J. Y. Tu, 2009: Evaluating the “Rich-Get-Richer’
mechanism in tropical precipitation change under global warming. J. Clim., 22,
1982–2005.
Chou, C., C. Chen, P.-H. Tan, and K.-T. Chen, 2012: Mechanisms for global warming
impacts on precipitation frequency and intensity. J. Clim., 25, 3291–3306.
Chou, C., J. C. H. Chiang, C.-W. Lan, C.-H. Chung, Y.-C. Liao, and C.-J. Lee, 2013:
Increase in the range between wet and dry season precipitation. Nature Geosci.,
6, 263–267.
Christensen, J. H., F. Boberg, O. B. Christensen, and P. Lucas-Picher, 2008: On the need
for bias correction of regional climate change projections of temperature and
precipitation. Geophys. Res. Lett., 35, L20709.
Christensen, J. H., et al., 2007: Regional climate projections. In: Climate Change
2007: The Physical Science Basis. Contribution of Working Group I to the
Fourth Assessment Report of the Intergovernmental Panel on Climate Change
[Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor
and H. L. Miller (eds.)] Cambridge University Press, Cambridge, United Kingdom
and New York, NY, USA, pp. 847–940.
Christensen, N., and D. Lettenmaier, 2007: A multimodel ensemble approach to
assessment of climate change impacts on the hydrology and water resources of
the Colorado River Basin. Hydrol. Earth Syst. Sci., 11, 1417–1434.
Cionni, I., et al., 2011: Ozone database in support of CMIP5 simulations: Results and
corresponding radiative forcing. Atmos. Chem. Phys., 11, 11267–11292.
Clark, R. T., S. J. Brown, and J. M. Murphy, 2006: Modeling Northern Hemisphere
summer heat extreme changes and their uncertainties using a physics ensemble
of climate sensitivity experiments. J. Clim., 19, 4418–4435.
Clark, R. T., J. M. Murphy, and S. J. Brown, 2010: Do global warming targets limit
heatwave risk? Geophys. Res. Lett., 37, L17703.
Claussen, M., V. Brovkin, A. Ganopolski, C. Kubatzki, and V. Petoukhov, 2003: Climate
change in northern Africa: The past is not the future. Clim. Change, 57, 99–118.
Colle, B. A., Z. Zhang, K. A. Lombardo, E. Chang, P. Liu, and M. Zhang, 2013: Historical
evaluation and future prediction of eastern North America and western Atlantic
extratropical cyclones in the CMIP5 models during the cool season. J. Clim.,
doi:10.1175/JCLI-D-12-00498.1.
Collier, J., and G. Zhang, 2009: Aerosol direct forcing of the summer Indian monsoon
as simulated by the NCAR CAM3. Clim. Dyn., 32, 313–332.
Collins, M., C. M. Brierley, M. MacVean, B. B. B. Booth, and G. R. Harris, 2007: The
sensitivity of the rate of transient climate change to ocean physics perturbations.
J. Clim., 20, 2315–2320.
Collins, M., B. B. B. Booth, G. Harris, J. M. Murphy, D. M. H. Sexton, and M. J. Webb,
2006a: Towards quantifying uncertainty in transient climate change. Clim. Dyn.,
27, 127–147.
Collins, M., R. E. Chandler, P. M. Cox, J. M. Huthnance, J. Rougier, and D. B. Stephenson,
2012: Quantifying future climate change. Nature Clim. Change, 2, 403–409.
Collins, M., B. Booth, B. Bhaskaran, G. Harris, J. Murphy, D. Sexton, and M. Webb,
2011: Climate model errors, feedbacks and forcings: A comparison of perturbed
physics and multi-model ensembles. Clim. Dyn., 36, 1737–1766.
Collins, M., et al., 2010: The impact of global warming on the tropical Pacific ocean
and El Nino. Nature Geosci., 3, 391–397.
Collins, W. D., et al., 2006b: Radiative forcing by well-mixed greenhouse gases:
Estimates from climate models in the Intergovernmental Panel on Climate
Change (IPCC) Fourth Assessment Report (AR4). J. Geophys. Res., 111, D14317.
Colman, R., and B. McAvaney, 2009: Climate feedbacks under a very broad range of
forcing. Geophys. Res. Lett., 36, L01702.
Colman, R., and S. Power, 2010: Atmospheric radiative feedbacks associated with
transient climate change and climate variability. Clim. Dyn., 34, 919–933.
Comiso, J. C., and F. Nishio, 2008: Trends in the sea ice cover using enhanced and
compatible AMSR-E, SSM/I, and SMMR data. J. Geophys. Res., 113, C02S07.
Cook, K., and E. Vizy, 2008: Effects of twenty-first-century climate change on the
Amazon rain forest. J. Clim., 21, 542–560.
Costa, M., and G. Pires, 2010: Effects of Amazon and Central Brazil deforestation
scenarios on the duration of the dry season in the arc of deforestation. Int. J.
Climatol., 30, 1970–1979.
Crook, J. A., P. M. Forster, and N. Stuber, 2011: Spatial patterns of modeled climate
feedback and contributions to temperature response and polar amplification. J.
Clim., 24, 3575–3592.
Crucifix, M., 2006: Does the Last Glacial Maximum constrain climate sensitivity?
Geophys. Res. Lett., 33, L18701.
1123
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Cruz, F. T., A. J. Pitman, J. L. McGregor, and J. P. Evans, 2010: Contrasting regional
responses to increasing leaf-level atmospheric carbon dioxide over Australia. J.
Hydrometeorol., 11, 296–314.
Dai, A., 2011: Drought under global warming: A review. WIREs Clim. Change, 2,
45–65.
Dai, A., 2013: Increasing drought under global warming in observations and models.
Nature Clim. Change, 3, 52–58.
Dakos, V., M. Scheffer, E. H. van Nes, V. Brovkin, V. Petoukhov, and H. Held, 2008:
Slowing down as an early warning signal for abrupt climate change. Proc. Natl.
Acad. Sci. U.S.A., 105, 14308–14312.
Danabasoglu, G., and P. Gent, 2009: Equilibrium climate sensitivity: Is it accurate to
use a slab ocean model? J. Clim., 22, 2494–2499.
Davin, E. L., N. de Noblet-Ducoudre, and P. Friedlingstein, 2007: Impact of land cover
change on surface climate: Relevance of the radiative forcing concept. Geophys.
Res. Lett., 34, L13702.
Davis, S., K. Caldeira, and H. Matthews, 2010: Future CO
2
emissions and climate
change from existing energy infrastructure. Science, 329, 1330–1333.
Davis, S. M., and K. H. Rosenlof, 2012: A multidiagnostic intercomparison of tropical-
width time series using reanalyses and satellite observations. J. Clim., 25, 1061–
1078.
De Angelis, H., and P. Skvarca, 2003: Glacier surge after ice shelf collapse. Science,
299, 1560–1562.
de Vries, H., R. J. Haarsma, and W. Hazeleger, 2012: Western European cold spells in
current and future climate. Geophys. Res. Lett., 39, L04706.
de Vries, P., and S. Weber, 2005: The Atlantic freshwater budget as a diagnostic for
the existence of a stable shut down of the meridional overturning circulation.
Geophys. Res. Lett., 32, L09606.
Del Genio, A. D., M.-S. Yao, and J. Jonas, 2007: Will moist convection be stronger in a
warmer climate? Geophys. Res. Lett., 34, L16703.
Delisle, G., 2007: Near-surface permafrost degradation: How severe during the 21st
century? Geophys. Res. Lett., 34, L09503.
Delworth, T. L., et al., 2008: The potential for abrupt change in the Atlantic meridional
overturning circulation. In: Abrupt Climate Change: A Report by the U.S. Climate
Change Science Program and the Subcommittee on Global Change Research,
U.S. Geological Survey, Washington, DC, pp. 258–359.
deMenocal, P., J. Ortiz, T. Guilderson, J. Adkins, M. Sarnthein, L. Baker, and M.
Yarusinsky, 2000: Abrupt onset and termination of the African Humid Period:
Rapid climate responses to gradual insolation forcing. Quaternary Science
Reviews, 19, 347–361.
Deser, C., A. Phillips, V. Bourdette, and H. Teng, 2012a: Uncertainty in climate change
projections: The role of internal variability. Clim. Dyn., 38, 527–546.
Deser, C., R. Knutti, S. Solomon, and A. S. Phillips, 2012b: Communication of the role
of natural variability in future North American climate. Nature Clim. Change, 2,
775–779.
Dessai, S., X. F. Lu, and M. Hulme, 2005: Limited sensitivity analysis of regional
climate change probabilities for the 21st century. J. Geophys. Res. Atmos., 110,
D19108.
Diffenbaugh, N. S., and M. Ashfaq, 2010: Intensification of hot extremes in the
United States. Geophys. Res. Lett., 37, L15701.
Diffenbaugh, N. S., J. S. Pal, F. Giorgi, and X. J. Gao, 2007: Heat stress intensification
in the Mediterranean climate change hotspot. Geophys. Res. Lett., 34, L11706.
Dijkstra, H., 2007: Characterization of the multiple equilibria regime in a global
ocean model. Tellus A, 59, 695–705.
DiNezio, P. N., A. C. Clement, G. A. Vecchi, B. J. Soden, and B. P. Kirtman, 2009: Climate
response of the equatorial Pacific to global warming. J. Clim., 22, 4873–4892.
Dirmeyer, P. A., Y. Jin, B. Singh, and X. Yan, 2013: Evolving land-atmosphere
interactions over North America from CMIP5 simulations. J. Clim., doi:10.1175/
JCLI-D-12-00454.1.
Dix, M., et al., 2013: The ACCESS Coupled Model: Documentation of core CMIP5
simulations and initial results. Aust. Meteorol. Oceanogr. J., 63, 83-199.
Dole, R., et al., 2011: Was there a basis for anticipating the 2010 Russian heat wave?
Geophys. Res. Lett., 38, L06702.
Dolman, A., G. van der Werf, M. van der Molen, G. Ganssen, J. Erisman, and B.
Strengers, 2010: A carbon cycle science update since IPCC AR-4. Ambio, 39,
402–412.
Donat, M. G., et al., 2013: Updated analyses of temperature and precipitation
extreme indices since the beginning of the twentieth century: The HadEX2
dataset. J. Geophys. Res., 118, 2098–2118.
Dong, B. W., J. M. Gregory, and R. T. Sutton, 2009: Understanding land-sea warming
contrast in response to increasing greenhouse gases. Part I: Transient adjustment.
J. Clim., 22, 3079–3097.
Dorrepaal, E., S. Toet, R. van Logtestijn, E. Swart, M. van de Weg, T. Callaghan, and
R. Aerts, 2009: Carbon respiration from subsurface peat accelerated by climate
warming in the subarctic. Nature, 460, 616–619.
scher, R., and T. Koenigk, 2013: Arctic rapid sea ice loss events in regional coupled
climate scenario experiments. Ocean Sci., 9, 217–248.
Doutriaux-Boucher, M., M. J. Webb, J. M. Gregory, and O. Boucher, 2009: Carbon
dioxide induced stomatal closure increases radiative forcing via a rapid
reduction in low cloud. Geophys. Res. Lett., 36, L02703.
Douville, H., J. F. Royer, J. Polcher, P. Cox, N. Gedney, D. B. Stephenson, and P. J. Valdes,
2000: Impact of CO
2
doubling on the Asian summer monsoon: Robust versus
model-dependent responses. J. Meteorol. Soc. Jpn., 78, 421–439.
Dowdy, A. J., G. A. Mills, B. Timbal, and Y. Wang, 2013: Changes in the risk of
extratropical cyclones in Eastern Australia. J. Clim., 26, 1403–1417.
Downes, S., A. Budnick, J. Sarmiento, and R. Farneti, 2011: Impacts of wind stress on
the Antarctic Circumpolar Current fronts and associated subduction. Geophys.
Res. Lett., 38, L11605.
Downes, S. M., and A. M. Hogg, 2013: Southern Ocean circulation and eddy
compensation in CMIP5 models. J. Clim., doi:10.1175/JCLI-D-12-00504.1.
Downes, S. M., N. L. Bindoff, and S. R. Rintoul, 2010: Changes in the subduction of
Southern Ocean water masses at the end of the twenty-first century in eight
IPCC models. J. Clim., 23, 6526–6541.
Driesschaert, E., et al., 2007: Modeling the influence of Greenland ice sheet melting
on the Atlantic meridional overturning circulation during the next millennia.
Geophys. Res. Lett., 34, L10707.
Drijfhout, S., G. J. van Oldenborgh, and A. Cimatoribus, 2012: Is a decline of AMOC
causing the warming hole above the North Atlantic in observed and modeled
warming patterns? J. Clim., 25, 8373–8379.
Drijfhout, S. S., S. Weber, and E. van der Swaluw, 2010: The stability of the MOC as
diagnosed from the model projections for the pre-industrial, present and future
climate. Clim. Dyn., 37, 1575–1586.
Dufresne, J.-L., et al., 2013: Climate change projections using the IPSL-CM5 Earth
system model: From CMIP3 to CMIP5. Clim. Dyn., 40, 2123–2165.
Dufresne, J., J. Quaas, O. Boucher, S. Denvil, and L. Fairhead, 2005: Contrasts in the
effects on climate of anthropogenic sulfate aerosols between the 20th and the
21st century. Geophys. Res. Lett., 32, L21703.
Dufresne, J. L., and S. Bony, 2008: An assessment of the primary sources of spread
of global warming estimates from coupled atmosphere-ocean models. J. Clim.,
21, 5135–5144.
Dulamsuren, C., M. Hauck, and M. Muhlenberg, 2008: Insect and small mammal
herbivores limit tree establishment in northern Mongolian steppe. Plant Ecol.,
195, 143–156.
Dulamsuren, C., M. Hauck, and C. Leuschner, 2010: Recent drought stress leads to
growth reductions in Larix sibirica in the western Khentey, Mongolia. Global
Change Biol., 16, 3024–3035.
Dulamsuren, C., et al., 2009: Water relations and photosynthetic performance in
Larix sibirica growing in the forest-steppe ecotone of northern Mongolia. Tree
Physiol., 29, 99–110.
Dunne, J. P., R. J. Stouffer, and J. G. John, 2013: Reductions in labour capacity
from heat stress under climate warming. Nature Clim. Change, doi:10.1038/
nclimate1827.
Durack, P., and S. Wijffels, 2010: Fifty-year trends in global ocean salinities and their
relationship to broad-scale warming. J. Clim., 23, 4342–4362.
Durack, P. J., S. E. Wijffels, and R. J. Matear, 2012: Ocean salinities reveal strong
global water cycle intensification during 1950 to 2000. Science, 336, 455–458.
Eby, M., K. Zickfeld, A. Montenegro, D. Archer, K. Meissner, and A. Weaver, 2009:
Lifetime of anthropogenic climate change: Millennial time scales of potential
CO
2
and surface temperature perturbations. J. Clim., 22, 2501–2511.
Edwards, T., M. Crucifix, and S. Harrison, 2007: Using the past to constrain the future:
How the palaeorecord can improve estimates of global warming. Prog. Phys.
Geogr., 31, 481–500.
Eglin, T., et al., 2010: Historical and future perspectives of global soil carbon response
to climate and land-use changes. Tellus B, 62, 700–718.
Eisenman, I., 2012: Factors controlling the bifurcation structure of sea ice retreat. J.
Geophys. Res., 117, D01111.
Eisenman, I., and J. Wettlaufer, 2009: Nonlinear threshold behavior during the loss of
Arctic sea ice. Proc. Natl. Acad. Sci. U.S.A., 106, 28–32.
1124
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Eisenman, I., T. Schneider, D. S. Battisti, and C. M. Bitz, 2011: Consistent changes in
the sea ice seasonal cycle in response to global warming. J. Clim., 24, 5325–
5335.
Eliseev, A., P. Demchenko, M. Arzhanov, and I. Mokhov, 2013: Transient hysteresis of
near-surface permafrost response to external forcing. Clim. Dyn., doi:10.1007/
s00382–013–1672–5.
Emori, S., and S. Brown, 2005: Dynamic and thermodynamic changes in mean and
extreme precipitation under changed climate. Geophys. Res. Lett., 32, L17706.
Eyring, V., et al., 2005: A strategy for process-oriented validation of coupled
chemistry-climate models. Bull. Am. Meteorol. Soc., 86, 1117–1133.
Eyring, V., et al., 2013: Long-term ozone changes and associated climate impacts in
CMIP5 simulations. J. Geophys. Res., doi:10.1002/jgrd.50316.
Falloon, P. D., R. Dankers, R. A. Betts, C. D. Jones, B. B. B. Booth, and F. H. Lambert,
2012: Role of vegetation change in future climate under the A1B scenario
and a climate stabilisation scenario, using the HadCM3C Earth system model.
Biogeosciences, 9, 4739–4756.
Farneti, R., and P. Gent, 2011: The effects of the eddy-induced advection coefficient
in a coarse-resolution coupled climate model. Ocean Model., 39, 135–145.
Farneti, R., T. Delworth, A. Rosati, S. Griffies, and F. Zeng, 2010: The role of mesoscale
eddies in the rectification of the Southern Ocean response to climate change. J.
Phys. Oceanogr., 40, 1539–1557.
Fasullo, J. T., 2010: Robust land-ocean contrasts in energy and water cycle feedbacks.
J. Clim., 23, 4677–4693.
Favre, A., and A. Gershunov, 2009: North Pacific cyclonic and anticyclonic transients
in a global warming context: Possible consequences for Western North American
daily precipitation and temperature extremes. Clim. Dyn., 32, 969–987.
Finnis, J., M. M. Holland, M. C. Serreze, and J. J. Cassano, 2007: Response of Northern
Hemisphere extratropical cyclone activity and associated precipitation to
climate change, as represented by the Community Climate System Model. J.
Geophys. Res., 112, G04S42.
Fischer, E. M., and C. Schär, 2009: Future changes in daily summer temperature
variability: Driving processes and role for temperature extremes. Clim. Dyn., 33,
917–935.
Fischer, E. M., and C. Schär, 2010: Consistent geographical patterns of changes in
high-impact European heatwaves. Nature Geosci., 3, 398–403.
Fischer, E. M., and R. Knutti, 2013: Robust projections of combined humidity and
temperature extremes. Nature Clim. Change, 3, 126–130.
Fischer, E. M., D. M. Lawrence, and B. M. Sanderson, 2011: Quantifying uncertainties
in projections of extremes—A perturbed land surface parameter experiment.
Clim. Dyn., 37, 1381–1398.
Fischer, E. M., J. Rajczak, and C. Schär, 2012a: Changes in European summer
temperature variability revisited. Geophys. Res. Lett., 39, L19702.
Fischer, E. M., K. W. Oleson, and D. M. Lawrence, 2012b: Contrasting urban and rural
heat stress responses to climate change. Geophys. Res. Lett., 39, L03705.
Flannery, B. P., 1984: Energy-balance models incorporating transport of thermal and
latent energy. J. Atmos. Sci., 41, 414–421.
Forest, C. E., P. H. Stone, and A. P. Sokolov, 2006: Estimated PDFs of climate system
properties including natural and anthropogenic forcings. Geophys. Res. Lett., 33,
L01705.
Forest, C. E., P. H. Stone, and A. P. Sokolov, 2008: Constraining climate model
parameters from observed 20th century changes. Tellus A, 60, 911–920.
Forster, P., and K. Taylor, 2006: Climate forcings and climate sensitivities diagnosed
from coupled climate model integrations. J. Clim., 19, 6181–6194.
Forster, P. M., T. Andrews, P. Good, J. M. Gregory, L. S. Jackson, and M. Zelinka, 2013:
Evaluating adjusted forcing and model spread for historical and future scenarios
in the CMIP5 generation of climate models. J. Geophys. Res., 118, 1139–1150.
Fowler, H., M. Ekstrom, S. Blenkinsop, and A. Smith, 2007a: Estimating change in
extreme European precipitation using a multimodel ensemble. J. Geophys. Res.,
112, D18104.
Fowler, H. J., S. Blenkinsop, and C. Tebaldi, 2007b: Linking climate change modelling
to impacts studies: Recent advances in downscaling techniques for hydrological
modelling. Int. J. Climatol., 27, 1547–1578.
Frame, D., B. Booth, J. Kettleborough, D. Stainforth, J. Gregory, M. Collins, and M.
Allen, 2005: Constraining climate forecasts: The role of prior assumptions.
Geophys. Res. Lett., 32, L09702.
Frederiksen, C. S., J. S. Frederiksen, J. M. Sisson, and S. L. Osbrough, 2011: Australian
winter circulation and rainfall changes and projections. Int. J. Clim. Change Strat.
Manage., 3, 170–188.
Friedlingstein, P., and S. Solomon, 2005: Contributions of past and present human
generations to committed warming caused by carbon dioxide. Proc. Natl. Acad.
Sci. U.S.A., 102, 10832–10836.
Friedlingstein, P., S. Solomon, G. Plattner, R. Knutti, P. Ciais, and M. Raupach, 2011:
Long-term climate implications of twenty-first century options for carbon
dioxide emission mitigation. Nature Clim. Change, 1, 457–461.
Friedlingstein, P., et al., 2006: Climate-carbon cycle feedback analysis: Results from
the C
4
MIP model intercomparison. J. Clim., 19, 3337–3353.
Frieler, K., M. Meinshausen, M. Mengel, N. Braun, and W. Hare, 2012: A scaling
approach to probabilistic assessment of regional climate. J. Clim., 25, 3117–
3144.
Frierson, D., J. Lu, and G. Chen, 2007: Width of the Hadley cell in simple and
comprehensive general circulation models. Geophys. Res. Lett., 34, L18804.
Frölicher, T., and F. Joos, 2010: Reversible and irreversible impacts of greenhouse gas
emissions in multi-century projections with the NCAR global coupled carbon
cycle-climate model. Clim. Dyn., 35, 1439–1459.
Fu, Q., C. M. Johanson, J. M. Wallace, and T. Reichler, 2006: Enhanced mid-latitude
tropospheric warming in satellite measurements. Science, 312, 1179–1179.
Fyfe, J., O. Saenko, K. Zickfeld, M. Eby, and A. Weaver, 2007: The role of poleward-
intensifying winds on Southern Ocean warming. J. Clim., 20, 5391–5400.
Fyke, J., and A. Weaver, 2006: The effect of potential future climate change on the
marine methane hydrate stability zone. J. Clim., 19, 5903–5917.
Garcia, R. R., and W. J. Randel, 2008: Acceleration of the Brewer-Dobson circulation
due to increases in greenhouse gases. J. Atmos. Sci., 65, 2731–2739.
Gastineau, G., and B. J. Soden, 2009: Model projected changes of extreme wind
events in response to global warming. Geophys. Res. Lett., 36, L10810.
Gastineau, G., H. Le Treut, and L. Li, 2008: Hadley circulation changes under global
warming conditions indicated by coupled climate models. Tellus A, 60, 863–884.
Gastineau, G., L. Li, and H. Le Treut, 2009: The Hadley and Walker circulation changes
in global warming conditions described by idealized atmospheric simulations. J.
Clim., 22, 3993–4013.
Gent, P. R., et al., 2011: The Community Climate System Model Version 4. J. Clim.,
24, 4973–4991.
Georgescu, M., D. Lobell, and C. Field, 2011: Direct climate effects of perennial
bioenergy crops in the United States. Proc. Natl. Acad. Sci. U.S.A., 109, 4307–
4312.
Gerber, E. P., et al., 2012: Assessing and understanding the impact of stratospheric
dynamics and variability on the Earth system. Bull. Am. Meteorol. Soc., 93,
845–859.
Gillett, N., M. Wehner, S. Tett, and A. Weaver, 2004: Testing the linearity of the
response to combined greenhouse gas and sulfate aerosol forcing. Geophys.
Res. Lett., 31, L14201.
Gillett, N. P., and P. A. Stott, 2009: Attribution of anthropogenic influence on seasonal
sea level pressure. Geophys. Res. Lett., 36, L23709.
Gillett, N. P., V. K. Arora, D. Matthews, and M. R. Allen, 2013: Constraining the ratio of
global warming to cumulative CO
2
emissions using CMIP5 simulations. J. Clim.,
doi:10.1175/JCLI-D-12-00476.1.
Gillett, N. P., V. K. Arora, K. Zickfeld, S. J. Marshall, and A. J. Merryfield, 2011: Ongoing
climate change following a complete cessation of carbon dioxide emissions.
Nature Geosci., 4, 83–87.
Giorgi, F., 2008: A simple equation for regional climate change and associated
uncertainty. J. Clim., 21, 1589–1604.
Gleckler, P. J., K. AchutaRao, J. M. Gregory, B. D. Santer, K. E. Taylor, and T. M. L. Wigley,
2006: Krakatoa lives: The effect of volcanic eruptions on ocean heat content and
thermal expansion. Geophys. Res. Lett., 33, L17702.
Goelzer, H., P. Huybrechts, M. Loutre, H. Goosse, T. Fichefet, and A. Mouchet, 2011:
Impact of Greenland and Antarctic ice sheet interactions on climate sensitivity.
Clim. Dyn., 37, 1005–1018.
Good, P., J. M. Gregory, and J. A. Lowe, 2011a: A step-response simple climate model
to reconstruct and interpret AOGCM projections. Geophys. Res. Lett., 38, L01703.
Good, P., J. M. Gregory, J. A. Lowe, and T. Andrews, 2013: Abrupt CO
2
experiments
as tools for predicting and understanding CMIP5 representative concentration
pathway projections. Clim. Dyn., 40, 1041–1053.
Good, P., C. Jones, J. Lowe, R. Betts, B. Booth, and C. Huntingford, 2011b: Quantifying
environmental drivers of future tropical forest extent. J. Clim., 24, 1337–1349.
Good, P., et al., 2012: A step-response approach for predicting and understanding
non-linear precipitation changes. Clim. Dyn., 39, 2789–2803.
1125
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Good, P., et al., 2011c: A review of recent developments in climate change science.
Part I: Understanding of future change in the large-scale climate system. Prog.
Phys. Geogr., 35, 281–296.
Goodwin, P., R. Williams, A. Ridgwell, and M. Follows, 2009: Climate sensitivity to the
carbon cycle modulated by past and future changes in ocean chemistry. Nature
Geosci., 2, 145–150.
Goosse, H., O. Arzel, C. Bitz, A. de Montety, and M. Vancoppenolle, 2009: Increased
variability of the Arctic summer ice extent in a warmer climate. Geophys. Res.
Lett., 36, L23702.
Goubanova, K., and L. Li, 2007: Extremes in temperature and precipitation around
the Mediterranean basin in an ensemble of future climate scenario simulations.
Global Planet. Change, 57, 27–42.
Gouttevin, I., G. Krinner, P. Ciais, J. Polcher, and C. Legout, 2012: Multi-scale validation
of a new soil freezing scheme for a land-surface model with physically-based
hydrology. Cryosphere, 6, 407–430.
Granier, C., et al., 2011: Evolution of anthropogenic and biomass burning emissions
at global and regional scales during the 1980–2010 period. Clim. Change, 109,
163–190.
Grant, A., S. Brönnimann, and L. Haimberger, 2008: Recent Arctic warming vertical
structure contested. Nature, 455, E2–E3.
Graversen, R., and M. Wang, 2009: Polar amplification in a coupled climate model
with locked albedo. Clim. Dyn., 33, 629–643.
Graversen, R., T. Mauritsen, M. Tjernstrom, E. Kallen, and G. Svensson, 2008: Vertical
structure of recent Arctic warming. Nature, 541, 53–56.
Gregory, J., and M. Webb, 2008: Tropospheric adjustment induces a cloud component
in CO
2
forcing. J. Clim., 21, 58–71.
Gregory, J., and P. Forster, 2008: Transient climate response estimated from radiative
forcing and observed temperature change. J. Geophys. Res., 113, D23105.
Gregory, J. M., 2010: Long-term effect of volcanic forcing on ocean heat content.
Geophys. Res. Lett., 37, L22701.
Gregory, J. M., and J. F. B. Mitchell, 1995: Simulation of daily variability of surface-
temperature and precipitation over Europe in the current and 2xCO
2
climates
using the UKMO climate model. Q. J. R. Meteorol. Soc., 121, 1451–1476.
Gregory, J. M., and R. Tailleux, 2011: Kinetic energy analysis of the response of the
Atlantic meridional overturning circulation to CO
2–
forced climate change. Clim.
Dyn., 37, 893–914.
Gregory, J. M., C. D. Jones, P. Cadule, and P. Friedlingstein, 2009: Quantifying carbon
cycle feedbacks. J. Clim., 22, 5232–5250.
Gregory, J. M., et al., 2004: A new method for diagnosing radiative forcing and
climate sensitivity. Geophys. Res. Lett., 31, L03205.
Gregory, J. M., et al., 2005: A model intercomparison of changes in the Atlantic
thermohaline circulation in response to increasing atmospheric CO
2
concentration. Geophys. Res. Lett., 32, L12703.
Grubb, M., 1997: Technologies, energy systems and the timing of CO
2
emissions
abatement—An overview of economic issues. Energy Policy, 25, 159–172.
Gumpenberger, M., et al., 2010: Predicting pan-tropical climate change induced
forest stock gains and losses-implications for REDD. Environ. Res. Lett., 5,
014013.
Gutowski, W., K. Kozak, R. Arritt, J. Christensen, J. Patton, and E. Takle, 2007: A
possible constraint on regional precipitation intensity changes under global
warming. J. Hydrometeorol., 8, 1382–1396.
Haarsma, R. J., F. Selten, and G. J. van Oldenborgh, 2013: Anthropogenic changes of
the thermal and zonal flow structure over Western Europe and Eastern North
Atlantic in CMIP3 and CMIP5 models. Clim. Dyn., doi:10.1007/s00382–013-
1734-8.
Haarsma, R. J., F. Selten, B. V. Hurk, W. Hazeleger, and X. L. Wang, 2009: Drier
Mediterranean soils due to greenhouse warming bring easterly winds over
summertime central Europe. Geophys. Res. Lett., 36, L04705.
Hajima, T., T. Ise, K. Tachiiri, E. Kato, S. Watanabe, and M. Kawamiya, 2012: Climate
change, allowable emission, and Earth system response to representative
concentration pathway scenarios. J. Meteorol. Soc. Jpn., 90, 417–433.
Hall, A., 2004: The role of surface albedo feedback in climate. J. Clim., 17, 1550–1568.
Hall, A., X. Qu, and J. Neelin, 2008: Improving predictions of summer climate change
in the United States. Geophys. Res. Lett., 35, L01702.
Hansen, J., M. Sato, P. Kharecha, and K. von Schuckmann, 2011: Earth’s energy
imbalance and implications. Atmos. Chem. Phys., 11, 13421–13449.
Hansen, J., G. Russell, A. Lacis, I. Fung, D. Rind, and P. Stone, 1985: Climate response-
times—Dependence on climate sensitivity and ocean mixing. Science, 229,
857–859.
Hansen, J., M. Sato, P. Kharecha, G. Russell, D. Lea, and M. Siddall, 2007: Climate
change and trace gases. Philos. Trans. R. Soc. A, 365, 1925–1954.
Hansen, J., et al., 1984: Climate sensitivity: Analysis of feedback mechanisms. In:
Climate Processes and Climate Sensitivity [J. Hansen and T. Takahashi (eds.)].
American Geophysical Union, Washington, DC, pp. 130–163.
Hansen, J., et al., 1988: Global climate changes as forecast by Goddard Institute for
Space Studies 3-dimensional model. J. Geophys. Res. Atmos., 93, 9341–9364.
Hansen, J., et al., 2008: Target atmospheric CO
2
: Where should humanity aim? Open
Atmos. Sci. J., 2, 217–231.
Hansen, J., et al., 2005a: Earth’s energy imbalance: Confirmation and implications.
Science, 308, 1431–1435.
Hansen, J., et al., 2005b: Efficacy of climate forcings. J. Geophys. Res., 110, D18104.
Hardiman, S., N. Butchart, T. Hinton, S. Osprey, and L. Gray, 2012: The effect of a
well resolved stratosphere on surface climate: Differences between CMIP5
simulations with high and low top versions of the Met Office climate model. J.
Clim., 35, 7083–7099.
Hare, B., and M. Meinshausen, 2006: How much warming are we committed to and
how much can be avoided? Clim. Change, 75, 111–149.
Hargreaves, J. C., A. Abe-Ouchi, and J. D. Annan, 2007: Linking glacial and future
climates through an ensemble of GCM simulations. Clim. Past, 3, 77–87.
Hargreaves, J. C., J. D. Annan, M. Yoshimori, and A. Abe-Ouchi, 2012: Can the Last
Glacial Maximum constrain climate sensitivity? Geophys. Res. Lett., 39, L24702.
Harris, G. R., M. Collins, D. M. H. Sexton, J. M. Murphy, and B. B. B. Booth, 2010:
Probabilistic projections for 21st century European climate. Nat. Hazards Earth
Syst. Sci., 10, 2009–2020.
Harris, G. R., D. M. H. Sexton, B. B. B. Booth, M. Collins, J. M. Murphy, and M. J.
Webb, 2006: Frequency distributions of transient regional climate change from
perturbed physics ensembles of general circulation model simulations. Clim.
Dyn., 27, 357–375.
Hartmann, D. L., and K. Larson, 2002: An important constraint on tropical cloud-
climate feedback. Geophys. Res. Lett., 29, 1951.
Harvey, B. J., L. C. Shaffrey, T. J. Woollings, G. Zappa, and K. I. Hodges, 2012: How
large are projected 21st century storm track changes? Geophys. Res. Lett., 39,
L18707.
Haugen, J., and T. Iversen, 2008: Response in extremes of daily precipitation and
wind from a downscaled multi-model ensemble of anthropogenic global climate
change scenarios. Tellus A, 60, 411–426.
Hawkins, E., and R. Sutton, 2009: The potential to narrow uncertainty in regional
climate predictions. Bull. Am. Meteorol. Soc., 90, 1095–1107.
Hawkins, E., and R. Sutton, 2011: The potential to narrow uncertainty in projections
of regional precipitation change. Clim. Dyn., 37, 407–418.
Hawkins, E., R. Smith, L. Allison, J. Gregory, T. Woollings, H. Pohlmann, and B. de
Cuevas, 2011: Bistability of the Atlantic overturning circulation in a global
climate model and links to ocean freshwater transport. Geophys. Res. Lett., 38,
L16699.
Hazeleger, W., et al., 2013: Multiyear climate predictions using two initialisation
strategies. Geophys. Res. Lett., doi:10.1002/grl.50355.
Hegerl, G., T. Crowley, W. Hyde, and D. Frame, 2006: Climate sensitivity constrained
by temperature reconstructions over the past seven centuries. Nature, 440,
1029–1032.
Hegerl, G. C., F. W. Zwiers, P. A. Stott, and V. V. Kharin, 2004: Detectability of
anthropogenic changes in annual temperature and precipitation extremes. J.
Clim., 17, 3683–3700.
Held, I., and B. Soden, 2006: Robust responses of the hydrological cycle to global
warming. J. Clim., 19, 5686–5699.
Held, I. M., M. Winton, K. Takahashi, T. Delworth, F. R. Zeng, and G. K. Vallis, 2010:
Probing the fast and slow components of global warming by returning abruptly
to preindustrial forcing. J. Clim., 23, 2418–2427.
Hellmer, H. H., F. Kauker, R. Timmermann, J. Determann, and J. Rae, 2012: Twenty-
first-century warming of a large Antarctic ice-shelf cavity by a redirected coastal
current. Nature, 484, 225–228.
Henderson-Sellers, A., P. Irannejad, and K. McGuffie, 2008: Future desertification
and climate change: The need for land-surface system evaluation improvement.
Global and Planetary Change, 64, 129–138.
Hibbard, K. A., G. A. Meehl, P. A. Cox, and P. Friedlingstein, 2007: A strategy for
climate change stabilization experiments. EOS Transactions AGU, 88, 217–221.
Hirschi, M., et al., 2011: Observational evidence for soil-moisture impact on hot
extremes in southeastern Europe. Nature Geosci., 4, 17–21.
1126
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Ho, C. K., D. B. Stephenson, M. Collins, C. A. T. Ferro, and S. J. Brown, 2012: Calibration
strategies: A source of additional uncertainty in climate change projections. Bull.
Am. Meteorol. Soc., 93, 21–26.
Hodson, D. L. R., S. P. E. Keeley, A. West, J. Ridley, E. Hawkins, and H. T. Hewitt,
2012: Identifying uncertainties in Arctic climate change projections. Clim. Dyn.,
doi:10.1007/s00382-012-1512-z.
Hoelzmann, P., D. Jolly, S. Harrison, F. Laarif, R. Bonnefille, and H. Pachur, 1998: Mid-
Holocene land-surface conditions in northern Africa and the Arabian Peninsula:
A data set for the analysis of biogeophysical feedbacks in the climate system.
Global Biogeochem. Cycles, 12, 35–51.
Hoerling, M., J. Eischeid, and J. Perlwitz, 2010: Regional precipitation trends:
Distinguishing natural variability from anthropogenic forcing. J. Clim., 23,
2131–2145.
Hoerling, M. P., J. K. Eischeid, X.-W. Quan, H. F. Diaz, R. S. Webb, R. M. Dole, and D. R.
Easterling, 2012: Is a transition to semipermanent drought conditions imminent
in the US Great Plains? J. Clim., 25, 8380–8386.
Hofmann, M., and S. Rahmstorf, 2009: On the stability of the Atlantic meridional
overturning circulation. Proc. Natl. Acad. Sci. U.S.A., 106, 20584–20589.
Hogg, E., and A. Schwarz, 1997: Regeneration of planted conifers across climatic
moisture gradients on the Canadian prairies: Implications for distribution and
climate change. J. Biogeogr., 24, 527–534.
Holden, P. B., and N. R. Edwards, 2010: Dimensionally reduced emulation of an
AOGCM for application to integrated assessment modelling. Geophys. Res. Lett.,
37, L21707.
Holland, M., C. Bitz, and B. Tremblay, 2006: Future abrupt reductions in the summer
Arctic sea ice. Geophys. Res. Lett., 33, L23503.
Holland, M., M. Serreze, and J. Stroeve, 2010: The sea ice mass budget of the Arctic
and its future change as simulated by coupled climate models. Clim. Dyn., 34,
185–200.
Holland, M. M., and C. M. Bitz, 2003: Polar amplification of climate change in
coupled models. Clim. Dyn., 21, 221–232.
Holland, M. M., C. M. Bitz, B. Tremblay, and D. A. Bailey, 2008: The role of natural
versus forced change in future rapid summer Arctic ice loss. In: Arctic Sea
Ice Decline: Observations, Projections, Mechanisms, and Implications [E. T.
DeWeaver, C. M. Bitz and L. B. Tremblay (eds.)]. American Geophysical Union,
Washington, DC, pp. 133–150.
Hu, A., G. Meehl, W. Han, and J. Yin, 2009: Transient response of the MOC and climate
to potential melting of the Greenland ice sheet in the 21st century. Geophys.
Res. Lett., 36, L10707.
Hu, Y., and Q. Fu, 2007: Observed poleward expansion of the Hadley circulation since
1979. Atmos. Chem. Phys., 7, 5229–5236.
Hu, Z.-Z., M. Latif, E. Roeckner, and L. Bengtsson, 2000: Intensified Asian summer
monsoon and its variability in a coupled model forced by increasing greenhouse
gas concentrations. Geophys. Res. Lett., 27, 2681–2684.
Hu, Z. Z., A. Kumar, B. Jha, and B. H. Huang, 2012: An analysis of forced and internal
variability in a warmer climate in CCSM3. J. Clim., 25, 2356–2373.
Huang, P., S.-P. Xie, K. Hu, G. Huang, and R. Huang, 2013: Patterns of the seasonal
response of tropical rainfall to global warming. Nature Geosci., 6, 357–361.
Huete, A. R., et al., 2006: Amazon rainforests green-up with sunlight in dry season.
Geophys. Res. Lett., 33, L06405.
Huisman, S., M. den Toom, H. Dijkstra, and S. Drijfhout, 2010: An indicator of the
multiple equilibria regime of the Atlantic meridional overturning circulation. J.
Phys. Oceanogr., 40, 551–567.
Huntingford, C., and P. M. Cox, 2000: An analogue model to derive additional climate
change scenarios from existing GCM simulations. Clim. Dyn., 16, 575–586.
Huntingford, C., J. Lowe, B. Booth, C. Jones, G. Harris, L. Gohar, and P. Meir, 2009:
Contributions of carbon cycle uncertainty to future climate projection spread.
Tellus B, 61, 355–360.
Huntingford, C., et al., 2008: Towards quantifying uncertainty in predictions of
Amazon ‘dieback’. Philos. Trans. R. Soc. B, 363, 1857–1864.
Huntingford, C., et al., 2013: Simulated resilience of tropical rainforests to CO
2–
induced climate change. Nature Geosci., 6, 268–273.
Hurtt, G., et al., 2011: Harmonization of land-use scenarios for the period 1500–
2100: 600 years of global gridded annual land-use transitions, wood harvest,
and resulting secondary lands. Clim. Change, 109, 117–161.
Hwang, Y.-T., D. M. W. D.M.W. Frierson, B. J. Soden, and I. M. Held, 2011: Corrigendum
for Held and Soden (2006). J. Clim., 24, 1559–1560.
IPCC, 2000: IPCC Special Report on Emissions Scenarios. Prepared by Working Group
III of the Intergovernmental Panel on Climate Change. Cambridge University
Press, Cambridge, United Kingdom, and New York, NY, USA.
IPCC, 2001: Climate Change 2001: The Scientific Basis. Contribution of Working
Group I to the Third Assessment Report of the Intergovernmental Panel on
Climate Change [J. T. Houghton, Y. Ding, D. J. Griggs, M. Noquer, P. J. van der
Linden, X. Dai, K. Maskell and C. A. Johnson (eds.)]. Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA, 881 pp.
IPCC, 2007: Climate Change 2007: The Physical Science Basis. Contribution of
Working Group I to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis,
K. B. Averyt, M. Tignor and H. L. Miller (eds.)]. Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA, 996 pp.
Ishizaki, Y., et al., 2012: Temperature scaling pattern dependence on representative
concentration pathway emission scenarios. Clim. Change, 112, 535–546.
Iversen, T., et al., 2013: The Norwegian Earth System Model, NorESM1–M – Part
2: Climate response and scenario projections. Geosci. Model Dev., 6, 389–415.
Jackson, C. S., M. K. Sen, G. Huerta, Y. Deng, and K. P. Bowman, 2008: Error reduction
and convergence in climate prediction. J. Clim., 21, 6698–6709.
Jaeger, C., and J. Jaeger, 2010: Three views of two degrees. Clim. Change Econ., 3,
145–166.
Jiang, X., S. J. Eichelberger, D. L. Hartmann, R. Shia, and Y. L. Yung, 2007: Influence of
doubled CO
2
on ozone via changes in the Brewer-Dobson circulation. J. Atmos.
Sci., 64, 2751–2755.
Johanson, C. M., and Q. Fu, 2009: Hadley Cell widening: Model simulations versus
observations. J. Clim., 22, 2713–2725.
Johns, T. C., et al., 2011: Climate change under aggressive mitigation: The ENSEMBLES
multi-model experiment. Clim. Dyn., 37, 1975–2003.
Johnson, N. C., and S.-P. Xie, 2010: Changes in the sea surface temperature threshold
for tropical convection. Nature Geosci., 3, 842–845.
Jones, A., J. Haywood, and O. Boucher, 2007: Aerosol forcing, climate response and
climate sensitivity in the Hadley Centre climate model. J. Geophys. Res., 112,
D20211.
Jones, C., P. Cox, and C. Huntingford, 2006: Climate-carbon cycle feedbacks under
stabilization: Uncertainty and observational constraints. Tellus B, 58, 603–613.
Jones, C., J. Lowe, S. Liddicoat, and R. Betts, 2009: Committed terrestrial ecosystem
changes due to climate change. Nature Geosci., 2, 484–487.
Jones, C. D., et al., 2013: 21st Century compatible CO
2
emissions and airborne
fraction simulated by CMIP5 Earth System models under 4 Representative
Concentration Pathways. J. Clim., doi:10.1175/JCLI-D-12-00554.1.
Jones, C. D., et al., 2011: The HadGEM2–ES implementation of CMIP5 centennial
simulations. Geosci. Model Dev., 4, 543–570.
Joos, F., et al., 2013: Carbon dioxide and climate impulse response functions for the
computation of greenhouse gas metrics: A multi-model analysis. Atmos. Chem.
Phys., 13, 2793–2825.
Joshi, M., E. Hawkins, R. Sutton, J. Lowe, and D. Frame, 2011: Projections of when
temperature change will exceed 2°C above pre-industrial levels. Nature Clim.
Change, 1, 407–412.
Joshi, M., K. Shine, M. Ponater, N. Stuber, R. Sausen, and L. Li, 2003: A comparison
of climate response to different radiative forcings in three general circulation
models: Towards an improved metric of climate change. Clim. Dyn., 20, 843–854.
Joshi, M. M., F. H. Lambert, and M. J. Webb, 2013: An explanation for the difference
between twentieth and twenty-first century land–sea warming ratio in climate
models. Clim. Dyn., doi:10.1007/s00382-013-1664-5.
Joshi, M. M., M. J. Webb, A. C. Maycock, and M. Collins, 2010: Stratospheric water
vapour and high climate sensitivity in a version of the HadSM3 climate model.
Atmos. Chem. Phys., 10, 7161–7167.
Joshi, M. M., J. M. Gregory, M. J. Webb, D. M. H. Sexton, and T. C. Johns, 2008:
Mechanisms for the land/sea warming contrast exhibited by simulations of
climate change. Clim. Dyn., 30, 455–465.
Jun, M., R. Knutti, and D. W. Nychka, 2008: Spatial analysis to quantify numerical
model bias and dependence: How many climate models are there? J. Am. Stat.
Assoc. Appl. Case Stud., 103, 934–947.
Jungclaus, J., H. Haak, M. Esch, E. Röckner, and J. Marotzke, 2006: Will Greenland
melting halt the thermohaline circulation? Geophys. Res. Lett., 33, L17708.
Kang, S. M., and I. M. Held, 2012: Tropical precipitation, SSTs and the surface energy
budget: A zonally symmetric perspective. Clim. Dyn., 38, 1917–1924.
Kang, S. M., L. M. Polvani, J. C. Fyfe, and M. Sigmond, 2011: Impact of polar ozone
depletion on subtropical precipitation. Science, 332, 951–954.
1127
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Karpechko, A. Y., and E. Manzini, 2012: Stratospheric influence on tropospheric
climate change in the Northern Hemisphere. J. Geophys. Res., 117, D05133.
Kattenberg, A., et al., 1996: Climate models—Projections of future climate. In:
Climate Change 1995: The Science of Climate Change. Contribution of WGI
to the Second Assessment Report of the Intergovernmental Panel on Climate
Change [J. T. Houghton, L. G. Meira . A. Callander, N. Harris, A. Kattenberg and
K. Maskell (eds.)]. Cambridge University Press, Cambridge, United Kingdom, and
New York, NY, USA, pp. 285–357.
Kawase, H., T. Nagashima, K. Sudo, and T. Nozawa, 2011: Future changes in
tropospheric ozone under Representative Concentration Pathways (RCPs).
Geophys. Res. Lett., 38, L05801.
Kay, J., M. Holland, and A. Jahn, 2011: Inter-annual to multi-decadal Arctic sea ice
extent trends in a warming world. Geophys. Res. Lett., 38, L15708.
Kay, J. E., M. M. Holland, C. Bitz, E. Blanchard-Wrigglesworth, A. Gettelman, A.
Conley, and D. Bailey, 2012: The influence of local feedbacks and northward heat
transport on the equilibrium Arctic climate response to increased greenhouse
gas forcing in coupled climate models. J. Clim., 25, 5433–5450.
Kaye, N., A. Hartley, and D. Hemming, 2012: Mapping the climate: Guidance on
appropriate techniques to map climate variables and their uncertainty. Geosci.
Model Dev., 5, 245–256.
Kellomaki, S., M. Maajarvi, H. Strandman, A. Kilpelainen, and H. Peltola, 2010: Model
computations on the climate change effects on snow cover, soil moisture and
soil frost in the boreal conditions over Finland. Silva Fennica, 44, 213–233.
Kendon, E., D. Rowell, and R. Jones, 2010: Mechanisms and reliability of future
projected changes in daily precipitation. Clim. Dyn., 35, 489–509.
Kendon, E., D. Rowell, R. Jones, and E. Buonomo, 2008: Robustness of future changes
in local precipitation extremes. J. Clim., 17, 4280–4297.
Kharin, V. V., F. W. Zwiers, X. B. Zhang, and G. C. Hegerl, 2007: Changes in temperature
and precipitation extremes in the IPCC ensemble of global coupled model
simulations. J. Clim., 20, 1419–1444.
Kharin, V. V., F. W. Zwiers, X. Zhang, and M. Wehner, 2013: Changes in temperature
and precipitation extremes in the CMIP5 ensemble. Clim. Change, doi:10.1007/
s10584-013-0705-8.
Khvorostyanov, D., P. Ciais, G. Krinner, and S. Zimov, 2008: Vulnerability of east
Siberia’s frozen carbon stores to future warming. Geophys. Res. Lett., 35, L10703.
Kidston, J., and E. P. Gerber, 2010: Intermodel variability of the poleward shift of the
austral jet stream in the CMIP3 integrations linked to biases in 20th century
climatology. Geophys. Res. Lett., 37, L09708.
Kienzle, S., M. Nemeth, J. Byrne, and R. MacDonald, 2012: Simulating the hydrological
impacts of climate change in the upper North Saskatchewan River basin, Alberta,
Canada. J. Hydrol., 412, 76–89.
Kirkevåg, K., et al., 2013: Aerosol–climate interactions in the Norwegian Earth
System Model – NorESM1–M. Geosci. Model Dev., 6, 207–244.
Kitoh, A., S. Yukimoto, A. Noda, and T. Motoi, 1997: Simulated changes in the Asian
summer monsoon at times of increased atmospheric CO
2
. J. Meteorol. Soc. Jpn.,
75, 1019–1031.
Kjellstrom, E., L. Barring, D. Jacob, R. Jones, G. Lenderink, and C. Schär, 2007:
Modelling daily temperature extremes: Recent climate and future changes over
Europe. Clim. Change, 81, 249–265.
Knutti, R., 2010: The end of model democracy? Clim. Change, 102, 395–404.
Knutti, R., and G. C. Hegerl, 2008: The equilibrium sensitivity of the Earth’s
temperature to radiation changes. Nature Geosci., 1, 735–743.
Knutti, R., and L. Tomassini, 2008: Constraints on the transient climate response
from observed global temperature and ocean heat uptake. Geophys. Res. Lett.,
35, L09701.
Knutti, R., and G.-K. Plattner, 2012: Comment on ‘Why hasn’t Earth warmed as much
as expected?’ by Schwartz et al. 2010. J. Clim., 25, 2192–2199.
Knutti, R., and J. Sedláček, 2013: Robustness and uncertainties in the new CMIP5
climate model projections. Nature Clim. Change, 3, 369–373.
Knutti, R., D. Masson, and A. Gettelman, 2013: Climate model genealogy: Generation
CMIP5 and how we got there. Geophys. Res. Lett., 40, 1194–1199.
Knutti, R., S. Krähenmann, D. Frame, and M. Allen, 2008a: Comment on “Heat
capacity, time constant, and sensitivity of Earth’s climate system’’ by S. E.
Schwartz. J. Geophys. Res., 113, D15103.
Knutti, R., F. Joos, S. Müller, G. Plattner, and T. Stocker, 2005: Probabilistic climate
change projections for CO
2
stabilization profiles. Geophys. Res. Lett., 32, L20707.
Knutti, R., R. Furrer, C. Tebaldi, J. Cermak, and G. A. Meehl, 2010a: Challenges in
combining projections from multiple climate models. J. Clim., 23, 2739–2758.
Knutti, R., G. Abramowitz, M. Collins, V. Eyring, P. J. Gleckler, B. Hewitson, and L.
Mearns, 2010b: Good practice guidance paper on assessing and combining
multi model climate projections. Meeting Report of the Intergovernmental Panel
on Climate Change Expert Meeting on Assessing and Combining Multi-Model
Climate Projections. IPCC Working Group I Technical Support Unit, University of
Bern, Bern, Switzerland.
Knutti, R., et al., 2008b: A review of uncertainties in global temperature projections
over the twenty-first century. J. Clim., 21, 2651–2663.
Kodra, E., K. Steinhaeuser, and A. R. Ganguly, 2011: Persisting cold extremes under
21st-century warming scenarios. Geophys. Res. Lett., 38, L08705.
Kolomyts, E., and N. Surova, 2010: Predicting the impact of global warming on soil
water resources in marginal forests of the middle Volga region. Water Resour.,
37, 89–101.
Komuro, Y., et al., 2012: Sea-ice in twentieth-century simulations by new MIROC
coupled models: A comparison between models with high resolution and with
ice thickness distribution. J. Meteorol. Soc. Jpn., 90A, 213–232.
rper, J., et al., 2013: The effects of aggressive mitigation on steric sea level rise and
sea ice changes. Clim. Dyn., 40, 531–550.
Koster, R., Z. Guo, R. Yang, P. Dirmeyer, K. Mitchell, and M. Puma, 2009a: On the
nature of soil moisture in land surface models. J. Clim., 22, 4322–4335.
Koster, R., et al., 2006: GLACE: The Global Land-Atmosphere Coupling Experiment.
Part I: Overview. J. Hydrometeorol., 7, 590–610.
Koster, R. D., S. D. Schubert, and M. J. Suarez, 2009b: Analyzing the concurrence
of meteorological droughts and warm periods, with implications for the
determination of evaporative regime. J. Clim., 22, 3331–3341.
Koster, R. D., H. L. Wang, S. D. Schubert, M. J. Suarez, and S. Mahanama, 2009c:
Drought-induced warming in the continental United States under different SST
regimes. J. Clim., 22, 5385–5400.
Koven, C., P. Friedlingstein, P. Ciais, D. Khvorostyanov, G. Krinner, and C. Tarnocai, 2009:
On the formation of high-latitude soil carbon stocks: Effects of cryoturbation
and insulation by organic matter in a land surface model. Geophys. Res. Lett.,
36, L21501.
Koven, C. D., W. J. Riley, and A. Stern, 2013: Analysis of permafrost thermal dynamics
and response to climate change in the CMIP5 Earth system models. J. Clim., 26,
1877–1900.
Koven, C. D., et al., 2011: Permafrost carbon-climate feedbacks accelerate global
warming. Proc. Natl. Acad. Sci. U.S.A., 108, 14769–14774.
Kripalani, R., J. Oh, A. Kulkarni, S. Sabade, and H. Chaudhari, 2007: South Asian
summer monsoon precipitation variability: Coupled climate model simulations
and projections under IPCC AR4. Theor. Appl. Climatol., 90, 133–159.
Kug, J., D. Choi, F. Jin, W. Kwon, and H. Ren, 2010: Role of synoptic eddy feedback
on polar climate responses to the anthropogenic forcing. Geophys. Res. Lett.,
37, L14704.
Kuhlbrodt, T., and J. M. Gregory, 2012: Ocean heat uptake and its consequences for
the magnitude of sea level rise and climate change. Geophys. Res. Lett., 39,
L18608.
Kuhry, P., E. Dorrepaal, G. Hugelius, E. Schuur, and C. Tarnocai, 2010: Potential
remobilization of belowground permafrost carbon under future global warming.
Permafr. Periglac. Process., 21, 208–214.
Kumar, A., et al., 2010: Contribution of sea ice loss to Arctic amplification. Geophys.
Res. Lett., 37, L21701.
Kunkel, K. E., T. R. Karl, D. R. Easterling, K. Redmond, J. Young, X. Yin, and P. Hennon,
2013: Probable Maximum Precipitation (PMP) and climate change. Geophys.
Res. Lett., 40, 1402–1408.
Kysely, J., and R. Beranova, 2009: Climate-change effects on extreme precipitation
in central Europe: Uncertainties of scenarios based on regional climate models.
Theor. Appl. Climatol., 95, 361–374.
Lamarque, J.-F., et al., 2011: Global and regional evolution of short-lived radiatively-
active gases and aerosols in the Representative Concentration Pathways. Clim.
Change, 109, 191–212.
Lamarque, J., 2008: Estimating the potential for methane clathrate instability in the
1%-CO
2
IPCC AR-4 simulations. Geophys. Res. Lett., 35, L19806.
Lamarque, J., et al., 2010: Historical (1850–2000) gridded anthropogenic and
biomass burning emissions of reactive gases and aerosols: Methodology and
application. Atmos. Chem. Phys., 10, 7017–7039.
Lamarque, J. F., et al., 2013: The Atmospheric Chemistry and Climate Model
Intercomparison Project (ACCMIP): Overview and description of models,
simulations and climate diagnostics. Geosci. Model Dev., 6, 179–206.
1128
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Lambert, F., and M. Webb, 2008: Dependency of global mean precipitation on surface
temperature. Geophys. Res. Lett., 35, L16706.
Lambert, F. H., and J. C. H. Chiang, 2007: Control of land-ocean temperature contrast
by ocean heat uptake. Geophys. Res. Lett., 34, L13704.
Lambert, F. H., M. J. Webb, and M. J. Joshi, 2011: The relationship between land-ocean
surface temperature contrast and radiative forcing. J. Clim., 24, 3239–3256.
Lambert, F. H., N. P. Gillett, D. A. Stone, and C. Huntingford, 2005: Attribution studies
of observed land precipitation changes with nine coupled models. Geophys. Res.
Lett., 32, L18704.
Lambert, F. H., G. R. Harris, M. Collins, J. M. Murphy, D. M. H. Sexton, and B. B. B. Booth,
2012: Interactions between perturbations to different Earth system components
simulated by a fully-coupled climate model. Clim. Dyn., doi:10.1007/s00382-
012-1618-3.
Langen, P. L., and V. A. Alexeev, 2007: Polar amplification as a preferred response in
an idealized aquaplanet GCM. Clim. Dyn., 29, 305–317.
Langen, P. L., A. M. Solgaard, and C. S. Hvidberg, 2012: Self-inhibiting growth of the
Greenland Ice Sheet. Geophys. Res. Lett., 39, L12502.
Lapola, D. M., M. D. Oyama, and C. A. Nobre, 2009: Exploring the range of climate
biome projections for tropical South America: The role of CO
2
fertilization and
seasonality. Global Biogeochem. Cycles, 23, GB3003.
Lau, K., M. Kim, and K. Kim, 2006: Asian summer monsoon anomalies induced by
aerosol direct forcing: The role of the Tibetan Plateau. Clim. Dyn., 26, 855–864.
Lawrence, D., and A. Slater, 2010: The contribution of snow condition trends to future
ground climate. Clim. Dyn., 34, 969–981.
Lawrence, D., A. Slater, and S. Swenson, 2012: Simulation of present-day and future
permafrost and seasonally frozen ground conditions in CCSM4. J. Clim., 25,
2207–2225.
Lawrence, D., A. Slater, V. Romanovsky, and D. Nicolsky, 2008a: Sensitivity of a
model projection of near-surface permafrost degradation to soil column depth
and representation of soil organic matter. J. Geophys. Res. Earth Surface, 113,
F02011.
Lawrence, D., A. Slater, R. Tomas, M. Holland, and C. Deser, 2008b: Accelerated Arctic
land warming and permafrost degradation during rapid sea ice loss. Geophys.
Res. Lett., 35, L11506.
Lean, J., and D. Rind, 2009: How will Earth’s surface temperature change in future
decades? Geophys. Res. Lett., 36, L15708.
Lee, S., T. Gong, N. Johnson, S. B. Feldstein, and D. Pollard, 2011: On the possible
link between tropical convection and the Northern Hemisphere Arctic surface air
temperature change between 1958 and 2001. J. Clim., 24, 4350–4367.
Lefebvre, W., and H. Goosse, 2008: Analysis of the projected regional sea-ice changes
in the Southern Ocean during the twenty-first century. Clim. Dyn., 30, 59–76.
Lemoine, D. M., 2010: Climate sensitivity distributions dependence on the possibility
that models share biases. J. Clim., 23, 4395–4415.
Lenderink, G., and E. Van Meijgaard, 2008: Increase in hourly precipitation extremes
beyond expectations from temperature changes. Nature Geosci., 1, 511–514.
Lenderink, G., A. van Ulden, B. van den Hurk, and E. van Meijgaard, 2007:
Summertime inter-annual temperature variability in an ensemble of regional
model simulations: Analysis of the surface energy budget. Clim. Change, 81,
233–247.
Lenton, T., H. Held, E. Kriegler, J. Hall, W. Lucht, S. Rahmstorf, and H. Schellnhuber,
2008: Tipping elements in the Earth’s climate system. Proc. Natl. Acad. Sci. U.S.A.,
105, 1786–1793.
Lenton, T. M., 2012: Arctic climate tipping points. Ambio, 41, 10–22.
Leslie, L. M., M. Leplastrier, and B. W. Buckley, 2008: Estimating future trends in
severe hailstorms over the Sydney Basin: A climate modelling study. Atmos. Res.,
87, 37–51.
Levermann, A., J. Schewe, V. Petoukhov, and H. Held, 2009: Basic mechanism for
abrupt monsoon transitions. Proc. Natl. Acad. Sci. U.S.A., 106, 20572–20577.
Levitus, S., J. Antonov, and T. Boyer, 2005: Warming of the world ocean, 1955–2003.
Geophys. Res. Lett., 32, L02604.
Levitus, S., et al., 2012: World ocean heat content and thermosteric sea level change
(0–2000 m), 1955–2010. Geophys. Res. Lett., 39, L10603.
Levy II, H., L. W. Horowitz, M. D. Schwarzkopf, Y. Ming, J.-C. Golaz, V. Naik, and V.
Ramaswamy, 2013: The roles of aerosol direct and indirect effects in past and
future climate change. J. Geophys. Res., doi:10.1002/jgrd.50192.
Li, C., J. S. von Storch, and J. Marotzke, 2013a: Deep-ocean heat uptake and
equilibrium climate response. Clim. Dyn., 40, 1071–1086.
Li, C., D. Notz, S. Tietsche, and J. Marotzke, 2013b: The transient versus the equilibrium
response of sea ice to global warming. J. Clim., doi:10.1175/JCLI-D-12-00492.1.
Li, F., J. Austin, and J. Wilson, 2008: The strength of the Brewer-Dobson circulation
in a changing climate: Coupled chemistry-climate model simulations. J. Clim.,
21, 40–57.
Li, F., W. Collins, M. Wehner, D. Williamson, J. Olson, and C. Algieri, 2011a: Impact of
horizontal resolution on simulation of precipitation extremes in an aqua-planet
version of Community Atmospheric Model (CAM3). Tellus, 63, 884–892.
Li, L., X. Jiang, M. Chahine, E. Olsen, E. Fetzer, L. Chen, and Y. Yung, 2011b: The
recycling rate of atmospheric moisture over the past two decades (1988–2009).
Environ. Res. Lett., 6, 034018.
Li, L. J., et al., 2013c: The Flexible Global Ocean-Atmosphere-Land System Model:
Grid-point Version 2: FGOALS-g2. Adv. Atmos. Sci., 30, 543–560.
Liepert, B. G., and M. Previdi, 2009: Do models and observations disagree on the
rainfall response to global warming? J. Clim., 22, 3156–3166.
Liepert, B. G., and M. Previdi, 2012: Inter-model variability and biases of the global
water cycle in CMIP3 coupled climate models. Environ. Res. Lett., 7, 014006.
Lim, E. P., and I. Simmonds, 2009: Effect of tropospheric temperature change on the
zonal mean circulation and SH winter extratropical cyclones. Clim. Dyn., 33,
19–32.
Lindsay, R., and J. Zhang, 2005: The thinning of Arctic sea ice, 1988–2003: Have we
passed a tipping point? J. Clim., 18, 4879–4894.
Liu, Z., S. J. Vavrus, F. He, N. Wen, and Y. Zhong, 2005: Rethinking tropical ocean
response to global warming: The enhanced equatorial warming. J. Clim., 18,
4684–4700.
Livina, V. N., and T. M. Lenton, 2013: A recent tipping point in the Arctic sea-ice cover:
Abrupt and persistent increase in the seasonal cycle since 2007. Cryosphere, 7,
275–286.
Loarie, S. R., D. B. Lobell, G. P. Asner, Q. Z. Mu, and C. B. Field, 2011: Direct impacts
on local climate of sugar-cane expansion in Brazil. Nature Clim. Change, 1,
105–109.
Loeb, N. G., et al., 2009: Toward optimal closure of the Earth’s Top-of-Atmosphere
radiation budget. J. Clim., 22, 748–766.
Long, M. C., K. Lindsay, S. Peacock, J. K. Moore, and S. C. Doney, 2013: Twentieth-
century oceanic carbon uptake and storage in CESM1(BGC). J. Clim., doi:10.1175/
JCLI-D-12-00184.1.
Lorenz, D. J., and E. T. DeWeaver, 2007: Tropopause height and zonal wind response
to global warming in the IPCC scenario integrations. J. Geophys. Res. Atmos.,
112, D10119.
Lowe, J., C. Huntingford, S. Raper, C. Jones, S. Liddicoat, and L. Gohar, 2009: How
difficult is it to recover from dangerous levels of global warming? Environ. Res.
Lett., 4, 014012.
Lowe, J. A., and J. M. Gregory, 2006: Understanding projections of sea level rise in a
Hadley Centre coupled climate model. J. Geophys. Res., 111, C11014.
Lu, J., and M. Cai, 2009: Seasonality of polar surface warming amplification in
climate simulations. Geophys. Res. Lett., 36, L16704.
Lu, J., G. Vecchi, and T. Reichler, 2007: Expansion of the Hadley cell under global
warming. Geophys. Res. Lett., 34, L06805.
Lu, J., G. Chen, and D. Frierson, 2008: Response of the zonal mean atmospheric
circulation to El Niño versus global warming. J. Clim., 21, 5835–5851.
Lucht, W., S. Schaphoff, T. Ebrecht, U. Heyder, and W. Cramer, 2006: Terrestrial
vegetation redistribution and carbon balance under climate change. Carbon
Balance Manage., 1, 1-6.
Lunt, D., A. Haywood, G. Schmidt, U. Salzmann, P. Valdes, and H. Dowsett, 2010:
Earth system sensitivity inferred from Pliocene modelling and data. Nature
Geosci., 3, 60–64.
Luo, J. J., W. Sasaki, and Y. Masumoto, 2012: Indian Ocean warming modulates Pacific
climate change. Proc. Natl. Acad. Sci. U.S.A., 109, 18701–18706.
Lustenberger, A., R. Knutti, and E. M. Fischer, 2013: The potential of pattern scaling
for projecting temperature-related extreme indices. Int. J. Climatol., doi:10.1002/
joc.3659.
Ma, J., and S.-P. Xie, 2013: Regional patterns of sea surface temperature change:
A source of uncertainty in future projections of precipitation and atmospheric
circulation. J. Clim., 26, 2482–2501.
Ma, J., S.-P. Xie, and Y. Kosaka, 2012: Mechanisms for tropical tropospheric circulation
change in response to global warming. J. Clim., 25, 2979–2994.
MacDougall, A. H., C. A. Avis, and A. J. Weaver, 2012: Significant contribution to
climate warming from the permafrost carbon feedback. Nature Geosci., 5,
719–721.
Mahlstein, I., and R. Knutti, 2011: Ocean heat transport as a cause for model
uncertainty in projected Arctic warming. J. Clim., 24, 1451–1460.
1129
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Mahlstein, I., and R. Knutti, 2012: September Arctic sea ice predicted to disappear
near 2°C global warming above present. J. Geophys. Res., 117, D06104.
Mahlstein, I., R. Knutti, S. Solomon, and R. W. Portmann, 2011: Early onset of
significant local warming in low latitude countries. Environ. Res. Lett., 6, 034009.
Mahlstein, I., R. W. Portmann, J. S. Daniel, S. Solomon, and R. Knutti, 2012: Perceptible
changes in regional precipitation in a future climate. Geophys. Res. Lett., 39,
L05701.
Maksym, T., S. E. Stammerjohn, S. Ackley, and R. Massom, 2012: Antarctic sea ice—A
polar opposite? Oceanography, 25, 140–151.
Malhi, Y., et al., 2009: Exploring the likelihood and mechanism of a climate-change-
induced dieback of the Amazon rainforest. Proc. Natl. Acad. Sci. U.S.A., 106,
20610–20615.
Manabe, S., and R. Stouffer, 1980: Sensitivity of a global climate model to an increase
of CO
2
concentration in the atmosphere. J. Geophys. Res., 85, 5529–5554.
Manabe, S., and R. T. Wetherald, 1980: Distribution of climate change resulting from
an increase in CO
2
content of the atmosphere. J. Atmos. Sci., 37, 99–118.
Manabe, S., and R. Stouffer, 1994: Multiple-century response of a coupled ocean-
atmosphere model to an increase of atmospheric carbon-dioxide. J. Clim., 7,
5–23.
Manabe, S., K. Bryan, and M. J. Spelman, 1990: Transient-response of a global
ocean atmosphere model to a doubling of atmospheric carbon-dioxide. J. Phys.
Oceanogr., 20, 722–749.
Manabe, S., R. J. Stouffer, M. J. Spelman, and K. Bryan, 1991: Transient responses of a
coupled ocean atmosphere model to gradual changes of atmospheric CO
2
. Part
I: Annual mean response. J. Clim., 4, 785–818.
Marsh, P. T., H. E. Brooks, and D. J. Karoly, 2009: Preliminary investigation into the
severe thunderstorm environment of Europe simulated by the Community
Climate System Model 3. Atmos. Res., 93, 607–618.
Maslowski, W., J. C. Kinney, M. Higgins, and A. Roberts, 2012: The future of Arctic sea
ice. In: Annual Review of Earth and Planetary Sciences [R. Jeanloz (ed.)]. Annual
Reviews, Palo Alto, CA, USA, pp. 625–654.
Masson, D., and R. Knutti, 2011: Climate model genealogy. Geophys. Res. Lett., 38,
L08703.
Massonnet, F., T. Fichefet, H. Goosse, C. M. Bitz, G. Philippon-Berthier, M. Holland,
and P. Y. Barriat, 2012: Constraining projections of summer Arctic sea ice.
Cryosphere, 6, 1383–1394.
Matsuno, T., K. Maruyama, and J. Tsutsui, 2012a: Stabilization of atmospheric carbon
dioxide via zero emissions-An alternative way to a stable global environment.
Part 1: Examination of the traditional stabilization concept. Proc. Jpn. Acad. B,
88, 368–384.
Matsuno, T., K. Maruyama, and J. Tsutsui, 2012b: Stabilization of atmospheric carbon
dioxide via zero emissions-An alternative way to a stable global environment.
Part 2: A practical zero-emissions scenario. Proc. Jpn. Acad. B, 88, 385–395.
Matthews, H., and K. Caldeira, 2008: Stabilizing climate requires near-zero emissions.
Geophys. Res. Lett., 35, L04705.
Matthews, H., N. Gillett, P. Stott, and K. Zickfeld, 2009: The proportionality of global
warming to cumulative carbon emissions. Nature, 459, 829–832.
Matthews, H. D., and K. Zickfeld, 2012: Climate response to zeroed emissions of
greenhouse gases and aerosols. Nature Clim. Change, 2, 338–341.
Matthews, H. D., S. Solomon, and R. Pierrehumbert, 2012: Cumulative carbon as
a policy framework for achieving climate stabilization. Philos. Trans. R. Soc. A,
370, 4365–4379.
May, W., 2002: Simulated changes of the Indian summer monsoon under enhanced
greenhouse gas conditions in a global time-slice experiment. Geophys. Res.
Lett., 29, 1118.
May, W., 2008a: Climatic changes associated with a global “2°C-stabilization”
scenario simulated by the ECHAM5/MPI-OM coupled climate model. Clim. Dyn.,
31, 283–313.
May, W., 2008b: Potential future changes in the characteristics of daily precipitation
in Europe simulated by the HIRHAM regional climate model. Clim. Dyn., 30,
581–603.
May, W., 2012: Assessing the strength of regional changes in near-surface climate
associated with a global warming of 2°C. Clim. Change, 110, 619–644.
McCabe, G., and D. Wolock, 2007: Warming may create substantial water supply
shortages in the Colorado River basin. Geophys. Res. Lett., 34, L22708.
McCollum, D., V. Krey, K. Riahi, P. Kolp, A. Grubler, M. Makowski, and N. Nakicenovic,
2013: Climate policies can help resolve energy security and air pollution
challenges. Clim. Change, doi:10.1007/s10584-013-0710-y.
McLandress, C., and T. G. Shepherd, 2009: Simulated anthropogenic changes in the
Brewer-Dobson circulation, including its extension to high latitudes. J. Clim., 22,
1516–1540.
McLandress, C., T. G. Shepherd, J. F. Scinocca, D. A. Plummer, M. Sigmond, A. I.
Jonsson, and M. C. Reader, 2011: Separating the dynamical effects of climate
change and ozone depletion. Part II: Southern Hemisphere troposphere. J. Clim.,
24, 1850–1868.
Meehl, G., and W. Washington, 1993: South Asian summer monsoon variability in a
model with doubled atmospheric carbon-dioxide concentration. Science, 260,
1101–1104.
Meehl, G., J. Arblaster, and C. Tebaldi, 2005a: Understanding future patterns of
increased precipitation intensity in climate model simulations. Geophys. Res.
Lett., 32, L18719.
Meehl, G., J. Arblaster, and W. Collins, 2008: Effects of black carbon aerosols on the
Indian monsoon. J. Clim., 21, 2869–2882.
Meehl, G., et al., 2012: Climate system response to external forcings and climate
change projections in CCSM4. J. Clim., 25, 3661–3683.
Meehl, G. A., and C. Tebaldi, 2004: More intense, more frequent, and longer lasting
heat waves in the 21st century. Science, 305, 994–997.
Meehl, G. A., G. J. Boer, C. Covey, M. Latif, and R. J. Stouffer, 2000: The Coupled Model
Intercomparison Project (CMIP). Bull. Am. Meteorol. Soc., 81, 313–318.
Meehl, G. A., C. Tebaldi, G. Walton, D. Easterling, and L. McDaniel, 2009: Relative
increase of record high maximum temperatures compared to record low
minimum temperatures in the U. S. Geophys. Res. Lett., 36, L23701.
Meehl, G. A., W. M. Washington, C. M. Ammann, J. M. Arblaster, T. M. L. Wigley, and C.
Tebaldi, 2004: Combinations of natural and anthropogenic forcings in twentieth-
century climate. J. Clim., 17, 3721–3727.
Meehl, G. A., et al., 2005b: How much more global warming and sea level rise?
Science, 307, 1769–1772.
Meehl, G. A., et al., 2007a: The WCRP CMIP3 multimodel dataset - A new era in
climate change research. Bull. Am. Meteorol. Soc., 88, 1383–1394.
Meehl, G. A., et al., 2006: Climate change projections for the twenty-first century and
climate change commitment in the CCSM3. J. Clim., 19, 2597–2616.
Meehl, G. A., et al., 2013: Climate change projections in CESM1(CAM5) compared to
CCSM4. J. Clim., doi:10.1175/JCLI-D-12–00572.1.
Meehl, G. A., et al., 2007b: Global climate projections. In: Climate Change 2007: The
Physical Science Basis. Contribution of Working Group I to the Fourth Assessment
Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin,
M. Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor and H. L. Miller (eds.)]
Cambridge University Press, Cambridge, United Kingdom and New York, NY,
USA, pp. 747–846.
Meijers, A. J. S., E. Shuckburgh, N. Bruneau, J.-B. Sallee, T. J. Bracegirdle, and Z. Wang,
2012: Representation of the Antarctic Circumpolar Current in the CMIP5 climate
models and future changes under warming scenarios. J. Geophys. Res., 117,
C12008.
Meinshausen, M., S. Raper, and T. Wigley, 2011a: Emulating coupled atmosphere-
ocean and carbon cycle models with a simpler model, MAGICC6–Part 1: Model
description and calibration. Atmos. Chem. Phys., 11, 1417–1456.
Meinshausen, M., T. Wigley, and S. Raper, 2011b: Emulating atmosphere-ocean
and carbon cycle models with a simpler model, MAGICC6–Part 2: Applications.
Atmos. Chem. Phys., 11, 1457–1471.
Meinshausen, M., B. Hare, T. Wigley, D. Van Vuuren, M. Den Elzen, and R. Swart,
2006: Multi-gas emissions pathways to meet climate targets. Clim. Change, 75,
151–194.
Meinshausen, M., et al., 2009: Greenhouse-gas emission targets for limiting global
warming to 2°C. Nature, 458, 1158–1162.
Meinshausen, M., et al., 2011c: The RCP greenhouse gas concentrations and their
extensions from 1765 to 2300. Clim. Change, 109, 213–241.
Merrifield, M. A., 2011: A shift in western tropical Pacific sea level trends during the
1990s. J. Clim., 24, 4126–4138.
Merryfield, W. J., M. M. Holland, and A. H. Monahan, 2008: Multiple equilibria and
abrupt transitions in Arctic summer sea ice extent. In: Arctic Sea Ice Decline:
Observations, Projections, Mechanisms, and Implications. American Geophysical
Union, Washington, DC, pp. 151–174.
Mignone, B., R. Socolow, J. Sarmiento, and M. Oppenheimer, 2008: Atmospheric
stabilization and the timing of carbon mitigation. Clim. Change, 88, 251–265.
Mikolajewicz, U., M. Vizcaino, J. Jungclaus, and G. Schurgers, 2007: Effect of ice sheet
interactions in anthropogenic climate change simulations. Geophys. Res. Lett.,
34, L18706.
1130
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Millner, A., R. Calel, D. A. Stainforth, and G. MacKerron, 2013: Do probabilistic expert
elicitations capture scientists’ uncertainty about climate change? Clim. Change,
116, 427–436.
Milly, P., J. Betancourt, M. Falkenmark, R. Hirsch, Z. Kundzewicz, D. Lettenmaier, and
R. Stouffer, 2008: Stationarity is dead: Whither water management? Science,
319, 573–574.
Min, S., X. Zhang, F. Zwiers, and G. Hegerl, 2011: Human contribution to more-
intense precipitation extremes. Nature, 470, 378–381.
Ming, Y., V. Ramaswamy, and G. Persad, 2010: Two opposing effects of absorbing
aerosols on global-mean precipitation. Geophys. Res. Lett., 37, L13701.
Mitas, C., and A. Clement, 2006: Recent behavior of the Hadley cell and tropical
thermodynamics in climate models and reanalyses. Geophys. Res. Lett., 33,
L01810.
Mitchell, J. F. B., 1990: Is the Holocene a good analogue for greenhouse warming?
J. Clim., 3, 1177–1192.
Mitchell, J. F. B., C. A. Wilson, and W. M. Cunnington, 1987: On CO
2
climate sensitivity
and model dependence of results. Q. J. R. Meteorol. Soc., 113, 293–322.
Mitchell, J. F. B., T. C. Johns, W. J. Ingram, and J. A. Lowe, 2000: The effect of stabilising
atmospheric carbon dioxide concentrations on global and regional climate
change. Geophys. Res. Lett., 27, 2977–2980.
Mitchell, J. F. B., T. C. Johns, M. Eagles, W. J. Ingram, and R. A. Davis, 1999: Towards the
construction of climate change scenarios. Clim. Change, 41, 547–581.
Mitchell, T. D., 2003: Pattern scaling - An examination of the accuracy of the
technique for describing future climates. Clim. Change, 60, 217–242.
Mizuta, R., 2012: Intensification of extratropical cyclones associated with the polar
jet change in the CMIP5 global warming projections. Geophys. Res. Lett., 39,
L19707.
Monaghan, A., D. Bromwich, and D. Schneider, 2008: Twentieth century Antarctic air
temperature and snowfall simulations by IPCC climate models. Geophys. Res.
Lett., 35, L07502.
Montenegro, A., V. Brovkin, M. Eby, D. Archer, and A. Weaver, 2007: Long term fate of
anthropogenic carbon. Geophys. Res. Lett., 34, L19707.
Morgan, M. G., and D. W. Keith, 1995: Climate-change - Subjective judgments by
climate experts. Environ. Sci. Technol., 29, A468–A476.
Moss, R. H., et al., 2010: The next generation of scenarios for climate change research
and assessment. Nature, 463, 747–756.
Moss, R. H., et al., 2008: Towards new scenarios for analysis of emissions, climate
change, impacts, and response strategies. In: IPCC Expert Meeting Report:
Towards New Scenarios. Intergovernmental Panel on Climate Change, Geneva,
Switzerland, 132 pp.
Muller, C. J., and P. A. O’Gorman, 2011: An energetic perspective on the regional
response of precipitation to climate change. Nature Clim. Change, 1, 266–271.
Murphy, D. M., S. Solomon, R. W. Portmann, K. H. Rosenlof, P. M. Forster, and T. Wong,
2009: An observationally based energy balance for the Earth since 1950. J.
Geophys. Res., 114, D17107.
Murphy, J., D. Sexton, D. Barnett, G. Jones, M. Webb, and M. Collins, 2004:
Quantification of modelling uncertainties in a large ensemble of climate change
simulations. Nature, 430, 768–772.
Murphy, J. M., B. B. B. Booth, M. Collins, G. R. Harris, D. M. H. Sexton, and M. J. Webb,
2007: A methodology for probabilistic predictions of regional climate change
from perturbed physics ensembles. Philos. Trans. R. Soc. A, 365, 1993–2028.
Myhre, G., E. Highwood, K. Shine, and F. Stordal, 1998: New estimates of radiative
forcing due to well mixed greenhouse gases. Geophys. Res. Lett., 25, 2715–2718.
Neelin, J. D., C. Chou, and H. Su, 2003: Tropical drought regions in global warming
and El Niño teleconnections. Geophys. Res. Lett., 30, 2275.
Neelin, J. D., M. Munnich, H. Su, J. E. Meyerson, and C. E. Holloway, 2006: Tropical
drying trends in global warming models and observations. Proc. Natl. Acad. Sci.
U.S.A., 103, 6110–6115.
Nelson, F., and S. Outcalt, 1987: A computational method for prediction and
regionalization of permafrost. Arct. Alpine Res., 19, 279–288.
Newlands, N. K., G. Espino-Hernández, and R. S. Erickson, 2012: Understanding crop
response to climate variability with complex agroecosystem models. Int. J. Ecol.,
2012, 756242.
Niall, S., and K. Walsh, 2005: The impact of climate change on hailstorms in
southeastern Australia. Int. J. Climatol., 25, 1933–1952.
Nicolsky, D., V. Romanovsky, V. Alexeev, and D. Lawrence, 2007: Improved modeling
of permafrost dynamics in a GCM land-surface scheme. Geophys. Res. Lett., 34,
L08501.
Niinemets, U., 2010: Responses of forest trees to single and multiple environmental
stresses from seedlings to mature plants: Past stress history, stress interactions,
tolerance and acclimation. Forest Ecol. Manage., 260, 1623–1639.
Nikulin, G., E. Kjellstrom, U. Hansson, G. Strandberg, and A. Ullerstig, 2011: Evaluation
and future projections of temperature, precipitation and wind extremes over
Europe in an ensemble of regional climate simulations. Tellus A, 63, 41–55.
Nobre, C., and L. Borma, 2009: ‘Tipping points’ for the Amazon forest. Curr. Opin.
Environ. Sustain., 1, 28–36.
North, G., 1984: The small ice cap instability in diffuse climate models. J. Atmos. Sci.,
41, 3390–3395.
Notaro, M., 2008: Statistical identification of global hot spots in soil moisture
feedbacks among IPCC AR4 models. J. Geophys. Res., 113, D09101.
Notz, D., 2009: The future of ice sheets and sea ice: Between reversible retreat and
unstoppable loss. Proc. Natl. Acad. Sci. U.S.A., 106, 20590–20595.
NRC, 2011: Climate Stabilization Targets: Emissions, Concentrations, and Impacts
over Decades to Millennia. National Academies Press, Washington, DC, 298 pp.
O’Connor, F., et al., 2010: Possible role of wetlands, permafrost, and methane
hydrates in the methane cycle under future climate change: A review. Rev.
Geophys., 48, RG4005.
O’Gorman, P., and T. Schneider, 2009a: Scaling of precipitation extremes over a wide
range of climates simulated with an idealized GCM. J. Clim., 22, 5676–5685.
O’Gorman, P., and T. Schneider, 2009b: The physical basis for increases in precipitation
extremes in simulations of 21st-century climate change. Proc. Natl. Acad. Sci.
U.S.A., 106, 14773–14777.
O’Gorman, P., R. Allan, M. Byrne, and M. Previdi, 2012: Energetic constraints on
precipitation under climate change. Surv. Geophys., 33, 585–608.
O’Gorman, P. A., 2010: Understanding the varied response of the extratropical storm
tracks to climate change. Proc. Natl. Acad. Sci. U.S.A., 107, 19176–19180.
O’Gorman, P. A., and C. J. Muller, 2010: How closely do changes in surface and
column water vapor follow Clausius-Clapeyron scaling in climate change
simulations? Environ. Res. Lett., 5, 025207.
Orlowsky, B., and S. I. Seneviratne, 2012: Global changes in extreme events: Regional
and seasonal dimension. Clim. Change, 110, 669–696.
Otto, A., et al., 2013: Energy budget constraints on climate response. Nature Geosci.,
6, 415-416.
Overland, J. E., and M. Wang, 2013: When will the summer Arctic be nearly sea ice
free? Geophys. Res. Lett., doi:10.1002/grl.50316.
Overland, J. E., M. Wang, N. A. Bond, J. E. Walsh, V. M. Kattsov, and W. L. Chapman,
2011: Considerations in the selection of global climate models for regional
climate projections: The Arctic as a case study. J. Clim., 24, 1583–1597.
Oyama, M. D., and C. A. Nobre, 2003: A new climate-vegetation equilibrium state for
Tropical South America. Geophys. Res. Lett., 30, 2199.
Padilla, L., G. Vallis, and C. Rowley, 2011: Probabilistic estimates of transient climate
sensitivity subject to uncertainty in forcing and natural variability. J. Clim., 24,
5521–5537.
Paeth, H., and F. Pollinger, 2010: Enhanced evidence in climate models for changes in
extratropical atmospheric circulation. Tellus A, 62, 647–660.
Pagani, M., Z. Liu, J. LaRiviere, and A. Ravelo, 2010: High Earth-system climate
sensitivity determined from Pliocene carbon dioxide concentrations. Nature
Geosci., 3, 27–30.
Pall, P., M. Allen, and D. Stone, 2007: Testing the Clausius-Clapeyron constraint on
changes in extreme precipitation under CO
2
warming. Clim. Dyn., 28, 351–363.
Pennell, C., and T. Reichler, 2011: On the effective number of climate models. J. Clim.,
24, 2358–2367.
Perkins, S. E., L. V. Alexander, and J. R. Nairn, 2012: Increasing frequency, intensity
and duration of observed global heatwaves and warm spells. Geophys. Res.
Lett., 39, L20714.
Perrie, W., Y. H. Yao, and W. Q. Zhang, 2010: On the impacts of climate change
and the upper ocean on midlatitude northwest Atlantic landfalling cyclones. J.
Geophys. Res., 115, D23110.
Piani, C., D. J. Frame, D. A. Stainforth, and M. R. Allen, 2005: Constraints on climate
change from a multi-thousand member ensemble of simulations. Geophys. Res.
Lett., 32, L23825.
Pierce, D., et al., 2008: Attribution of declining Western US snowpack to human
effects. J. Clim., 21, 6425–6444.
Pinto, J. G., U. Ulbrich, G. C. Leckebusch, T. Spangehl, M. Reyers, and S. Zacharias,
2007: Changes in storm track and cyclone activity in three SRES ensemble
experiments with the ECHAM5/MPI-OM1 GCM. Clim. Dyn., 29, 195–210.
1131
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Pitman, A., et al., 2009: Uncertainties in climate responses to past land cover
change: First results from the LUCID intercomparison study. Geophys. Res. Lett.,
36, L14814.
Plattner, G.-K., et al., 2008: Long-term climate commitments projected with climate-
carbon cycle models. J. Clim., 21, 2721– 2751.
Polvani, L. M., M. Previdi, and C. Deser, 2011: Large cancellation, due to ozone
recovery, of future Southern Hemisphere atmospheric circulation trends.
Geophys. Res. Lett., 38, L04707.
Pongratz, J., C. Reick, T. Raddatz, and M. Claussen, 2010: Biogeophysical versus
biogeochemical climate response to historical anthropogenic land cover change.
Geophys. Res. Lett., 37, L08702.
Port, U., V. Brovkin, and M. Claussen, 2012: The influence of vegetation dynamics on
anthropogenic climate change. Earth Syst. Dyn., 3, 233–243.
Power, S., and G. Kociuba, 2011a: The impact of global warming on the Southern
Oscillation Index. Clim. Dyn., 37, 1745–1754.
Power, S., F. Delage, R. Colman, and A. Moise, 2012: Consensus on twenty-first-
century rainfall projections in climate models more widespread than previously
thought. J. Clim., 25, 3792–3809.
Power, S. B., and G. Kociuba, 2011b: What caused the observed twentieth-century
weakening of the Walker circulation? J. Clim., 24, 6501–6514.
Previdi, M., 2010: Radiative feedbacks on global precipitation. Environ. Res. Lett.,
5, 025211.
Rahmstorf, S., et al., 2005: Thermohaline circulation hysteresis: A model
intercomparison. Geophys. Res. Lett., 32, L23605.
Räisänen, J., 2007: How reliable are climate models? Tellus A, 59, 2–29.
Räisänen, J., 2008: Warmer climate: Less or more snow? Clim. Dyn., 30, 307–319.
Räisänen, J., and L. Ruokolainen, 2006: Probabilistic forecasts of near-term climate
change based on a resampling ensemble technique. Tellus A, 58, 461–472.
Räisänen, J., and J. S. Ylhaisi, 2011: Cold months in a warming climate. Geophys. Res.
Lett., 38, L22704.
Ramanathan, V., P. J. Crutzen, J. T. Kiehl, and D. Rosenfeld, 2001: Aerosols, climate,
and the hydrologic cycle. Science, 294, 2119–2124.
Ramaswamy, V., et al., 2001: Radiative forcing of climate change. In: Climate
Change 2001: The Scientific Basis. Contribution of Working Group I to the Third
Assessment Report of the Intergovernmental Panel on Climate Change [J. T.
Houghton, Y. Ding, D. J. Griggs, M. Noquer, P. J. van der Linden, X. Dai, K. Maskell
and C. A. Johnson (eds.)]. Cambridge University Press, Cambridge, United
Kingdom and New York, NY, USA pp. 349-416.
Rammig, A., et al., 2010: Estimating the risk of Amazonian forest dieback. New
Phytologist, 187, 694–706.
Randall, D. A., et al., 2007: Climate models and their evaluation. In: Climate Change
2007: The Physical Science Basis. Contribution of Working Group I to the
Fourth Assessment Report of the Intergovernmental Panel on Climate Change
[Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor
and H. L. Miller (eds.)] Cambridge University Press, Cambridge, United Kingdom
and New York, NY, USA, pp. 589–662.
Randalls, S., 2010: History of the 2°C climate target. WIREs Climate Change, 1,
598–605.
Randel, W., and F. Wu, 2007: A stratospheric ozone profile data set for 1979–2005:
Variability, trends, and comparisons with column ozone data. J. Geophys. Res.,
112, D06313.
Randel, W. J., M. Park, F. Wu, and N. Livesey, 2007: A large annual cycle in ozone
above the tropical tropopause linked to the Brewer-Dobson circulation. J. Atmos.
Sci., 64, 4479–4488.
Randles, C., and V. Ramaswamy, 2008: Absorbing aerosols over Asia: A Geophysical
Fluid Dynamics Laboratory general circulation model sensitivity study of model
response to aerosol optical depth and aerosol absorption. J. Geophys. Res., 113,
D21203.
Reagan, M., and G. Moridis, 2007: Oceanic gas hydrate instability and dissociation
under climate change scenarios. Geophys. Res. Lett., 34, L22709.
Reagan, M., and G. Moridis, 2009: Large-scale simulation of methane hydrate
dissociation along the West Spitsbergen Margin. Geophys. Res. Lett., 36, L23612.
Ridley, J., J. Lowe, and D. Simonin, 2008: The demise of Arctic sea ice during
stabilisation at high greenhouse gas concentrations. Clim. Dyn., 30, 333–341.
Ridley, J., J. Lowe, C. Brierley, and G. Harris, 2007: Uncertainty in the sensitivity
of Arctic sea ice to global warming in a perturbed parameter climate model
ensemble. Geophys. Res. Lett., 34, L19704.
Ridley, J., J. Gregory, P. Huybrechts, and J. Lowe, 2010: Thresholds for irreversible
decline of the Greenland ice sheet. Clim. Dyn., 35, 1049–1057.
Ridley, J. K., J. A. Lowe, and H. T. Hewitt, 2012: How reversible is sea ice loss?
Cryosphere, 6, 193–198.
Riley, W. J., et al., 2011: Barriers to predicting changes in global terrestrial methane
fluxes: Analyses using CLM4ME, a methane biogeochemistry model integrated
in CESM. Biogeosciences, 8, 1925–1953.
Rind, D., 1987: The doubled CO
2
climate -Impact of the sea-surface temperature-
gradient. J. Atmos. Sci., 44, 3235–3268.
Rinke, A., P. Kuhry, and K. Dethloff, 2008: Importance of a soil organic layer for Arctic
climate: A sensitivity study with an Arctic RCM. Geophys. Res. Lett., 35, L13709.
Rive, N., A. Torvanger, T. Berntsen, and S. Kallbekken, 2007: To what extent can a
long-term temperature target guide near-term climate change commitments?
Clim. Change, 82, 373–391.
Robinson, A., R. Calov, and A. Ganopolski, 2012: Multistability and critical thresholds
of the Greenland ice sheet. Nature Clim. Change, 2, 429–432.
Roeckner, E., M. A. Giorgetta, T. Crueger, M. Esch, and J. Pongratz, 2011: Historical
and future anthropogenic emission pathways derived from coupled climate-
carbon cycle simulations. Clim. Change, 105, 91–108.
Roehrig, R., D. Bouniol, F. Guichard, F. Hourdin, and J.-L. Redelsperger, 2013: The
present and future of the West African monsoon: A process-oriented assessment
of CMIP5 simulations along the AMMA transect. J. Clim., doi:10.1175/JCLI-D-
12-00505.1.
Roesch, A., 2006: Evaluation of surface albedo and snow cover in AR4 coupled
climate models. J. Geophys. Res., 111, D15111.
Rogelj, J., M. Meinshausen, and R. Knutti, 2012: Global warming under old and new
scenarios using IPCC climate sensitivity range estimates. Nature Clim. Change,
2, 248–253.
Rogelj, J., D. L. McCollum, B. C. O’Neill, and K. Riahi, 2013: 2020 emissions levels
required to limit warming to below 2°C. Nature Clim. Change, 3, 405–412.
Rogelj, J., et al., 2011: Emission pathways consistent with a 2°C global temperature
limit. Nature Clim. Change, 1, 413–418.
Rohling, E., and P. P. Members, 2012: Making sense of palaeoclimate sensitivity.
Nature, 491, 683–691.
Rohling, E., K. Grant, M. Bolshaw, A. Roberts, M. Siddall, C. Hemleben, and M.
Kucera, 2009: Antarctic temperature and global sea level closely coupled over
the past five glacial cycles. Nature Geosci., 2, 500–504.
Romanovsky, V. E., S. L. Smith, and H. H. Christiansen, 2010: Permafrost thermal state
in the polar Northern Hemisphere during the international polar year 2007–
2009: A synthesis. Permafr. Periglac. Process., 21, 106–116.
Rotstayn, L. D., S. J. Jeffrey, M. A. Collier, S. M. Dravitzki, A. C. Hirst, J. I. Syktus, and
K. K. Wong, 2012: Aerosol- and greenhouse gas-induced changes in summer
rainfall and circulation in the Australasian region: A study using single-forcing
climate simulations. Atmos. Chem. Phys., 12, 6377–6404.
Rougier, J., 2007: Probabilistic inference for future climate using an ensemble of
climate model evaluations. Clim. Change, 81, 247–264.
Rougier, J., D. M. H. Sexton, J. M. Murphy, and D. Stainforth, 2009: Analyzing the
climate sensitivity of the HadSM3 climate model using ensembles from different
but related experiments. J. Clim., 22, 3540–3557.
Rowell, D. P., 2012: Sources of uncertainty in future changes in local precipitation.
Clim. Dyn., doi:10.1007/s00382–011–1210–2.
Rowlands, D. J., et al., 2012: Broad range of 2050 warming from an observationally
constrained large climate model ensemble. Nature Geosci., 5, 256–260.
Ruosteenoja, K., H. Tuomenvirta, and K. Jylha, 2007: GCM-based regional
temperature and precipitation change estimates for Europe under four SRES
scenarios applying a super-ensemble pattern-scaling method. Clim. Change, 81,
193–208.
Saenko, O. A., A. S. Gupta, and P. Spence, 2012: On challenges in predicting bottom
water transport in the Southern Ocean. J. Clim., 25, 1349–1356.
Saito, K., M. Kimoto, T. Zhang, K. Takata, and S. Emori, 2007: Evaluating a high-
resolution climate model: Simulated hydrothermal regimes in frozen ground
regions and their change under the global warming scenario. J. Geophys. Res.,
112, F02S11.
Sallée, J.-B., E. Shuckburgh, N. Bruneau, A. J. S. Meijers, T. Bracegirdle, and Z. Wang,
2013a: Assessment of Southern Ocean mixed-layer depths in CMIP5 models:
Historical bias and forcing response. J. Geophys. Res., doi:10.1002/jgrc.20157.
Sallée, J.-B., E. Shuckburgh, N. Bruneau, A. J. S. Meijers, T. J. Bracegirdle, Z. Wang,
and T. Roy, 2013b: Assessment of Southern Ocean water mass circulation and
characteristics in CMIP5 models: Historical bias and forcing response. J. Geophys.
Res., doi:10.1002/jgrc.20135.
1132
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Sanchez-Gomez, E., S. Somot, and A. Mariotti, 2009: Future changes in the
Mediterranean water budget projected by an ensemble of regional climate
models. Geophys. Res. Lett., 36, L21401.
Sanderson, B. M., 2011: A multimodel study of parametric uncertainty in predictions
of climate response to rising greenhouse gas concentrations. J. Clim., 25, 1362–
1377.
Sanderson, B. M., 2013: On the estimation of systematic error in regression-based
predictions of climate sensitivity. Clim. Change, doi:10.1007/s10584–012–
0671–6.
Sanderson, B. M., and R. Knutti, 2012: On the interpretation of constrained climate
model ensembles. Geophys. Res. Lett., 39, L16708.
Sanderson, B. M., K. M. Shell, and W. Ingram, 2010: Climate feedbacks determined
using radiative kernels in a multi-thousand member ensemble of AOGCMs. Clim.
Dyn., 35, 1219–1236.
Sanderson, B. M., et al., 2008: Constraints on model response to greenhouse gas
forcing and the role of subgrid-scale processes. J. Clim., 21, 2384–2400.
Sanderson, M. G., D. L. Hemming, and R. A. Betts, 2011: Regional temperature and
precipitation changes under high-end (>= 4°C) global warming. Philos. Trans. R.
Soc. A, 369, 85–98.
Sanso, B., and C. Forest, 2009: Statistical calibration of climate system properties. J.
R. Stat. Soc. C, 58, 485–503.
Sanso, B., C. E. Forest, and D. Zantedeschi, 2008: Inferring climate system properties
using a computer model. Bayes. Anal., 3, 1–37.
Sansom, P. G., D. B. Stephenson, C. A. T. Ferro, G. Zappa, and L. Shaffrey, 2013: Simple
uncertainty frameworks for selecting weighting schemes and interpreting multi-
model ensemble climate change experiments. J. Clim., doi:10.1175/JCLI-D-12-
00462.1.
Santer, B. D., T. M. L. Wigley, M. E. Schlesinger, and J. F. B. Mitchell, 1990: Developing
Climate Scenarios from Equilibrium GCM Results. Max-Planck-Institut-
r-Meteorologie Report. Max-Planck-Institut-für-Meteorologie, Hamburg,
Germany, 29 pp.
Scaife, A. A., et al., 2012: Climate change projections and stratosphere-troposphere
interaction. Clim. Dyn., 38, 2089–2097.
Schaefer, K., T. Zhang, L. Bruhwiler, and A. Barrett, 2011: Amount and timing of
permafrost carbon release in response to climate warming. Tellus B, 63, 165–
180.
Schär, C., P. L. Vidale, D. Lüthi, C. Frei, C. Häberli, M. A. Liniger, and C. Appenzeller,
2004: The role of increasing temperature variability in European summer
heatwaves. Nature, 427, 332–336.
Scheff, J., and D. M. W. Frierson, 2012: Robust future precipitation declines in CMIP5
largely reflect the poleward expansion of model subtropical dry zones. Geophys.
Res. Lett., 39, L18704.
Scheffer, M., et al., 2009: Early-warning signals for critical transitions. Nature, 461,
53–59.
Schlesinger, M., 1986: Equilibrium and transient climatic warming induced by
increased atmospheric CO
2
. Clim. Dyn., 1, 35–51.
Schlesinger, M., et al., 2000: Geographical distributions of temperature change for
scenarios of greenhouse gas and sulfur dioxide emissions. Technol. Forecast. Soc.
Change, 65, 167–193.
Schmidt, M. W. I., et al., 2011: Persistence of soil organic matter as an ecosystem
property. Nature, 478, 49–56.
Schmittner, A., et al., 2011: Climate sensitivity estimated from temperature
reconstructions of the Last Glacial Maximum. Science, 334, 1385–1388.
Schneider von Deimling, T., H. Held, A. Ganopolski, and S. Rahmstorf, 2006: Climate
sensitivity estimated from ensemble simulations of glacial climate. Clim. Dyn.,
27, 149–163.
Schneider von Deimling, T., M. Meinshausen, A. Levermann, V. Huber, K. Frieler, D.
Lawrence, and V. Brovkin, 2012: Estimating the near-surface permafrost-carbon
feedback on global warming. Biogeosciences, 9, 649–665.
Schoof, C., 2007: Ice sheet grounding line dynamics: Steady states, stability, and
hysteresis. J. Geophys. Res., 112, F03S28.
Schröder, D., and W. M. Connolley, 2007: Impact of instantaneous sea ice removal in
a coupled general circulation model. Geophys. Res. Lett., 34, L14502.
Schuenemann, K. C., and J. J. Cassano, 2010: Changes in synoptic weather patterns
and Greenland precipitation in the 20th and 21st centuries: 2. Analysis of 21st
century atmospheric changes using self-organizing maps. J. Geophys. Res., 115,
D05108.
Schuur, E., J. Vogel, K. Crummer, H. Lee, J. Sickman, and T. Osterkamp, 2009: The effect
of permafrost thaw on old carbon release and net carbon exchange from tundra.
Nature, 459, 556–559.
Schwalm, C. R., et al., 2012: Reduction in carbon uptake during turn of the century
drought in western North America. Nature Geosci., 5, 551–556.
Schwartz, S., R. Charlson, R. Kahn, J. Ogren, and H. Rodhe, 2010: Why hasn’t Earth
warmed as much as expected? J. Clim., 23, 2453–2464.
Schwartz, S., R. Charlson, R. Kahn, J. Ogren, and H. Rodhe, 2012: Reply to “Comments
on ‘Why hasn’t Earth warmed as much as expected?’”. J. Clim., 25, 2200–2204.
Schwartz, S. E., 2012: Determination of Earth’s transient and equilibrium climate
sensitivities from observations over the twentieth century: Strong dependence
on assumed forcing. Surv. Geophys., 33, 745–777.
Schweiger, A., R. Lindsay, J. Zhang, M. Steele, H. Stern, and R. Kwok, 2011: Uncertainty
in modeled Arctic sea ice volume. J. Geophys. Res., 116, C00D06.
Screen, J., and I. Simmonds, 2010: The central role of diminishing sea ice in recent
Arctic temperature amplification. Nature, 464, 1334–1337.
Screen, J. A., N. P. Gillett, A. Y. Karpechko, and D. P. Stevens, 2010: Mixed layer
temperature response to the Southern Annular Mode: Mechanisms and model
representation. J. Clim., 23, 664–678.
Seager, R., and G. A. Vecchi, 2010: Greenhouse warming and the 21st century
hydroclimate of the southwestern North America. Proc. Natl. Acad. Sci. U.S.A.,
107, 21277–21282.
Seager, R., and N. Naik, 2012: A mechanisms-based approach to detecting recent
anthropogenic hydroclimate change. J. Clim., 25, 236–261.
Seager, R., N. Naik, and G. A. Vecchi, 2010: Thermodynamic and dynamic mechanisms
for large-scale changes in the hydrological cycle in response to global warming.
J. Clim., 23, 4651–4668.
Seager, R., et al., 2007: Model projections of an imminent transition to a more arid
climate in southwestern North America. Science, 316, 1181–1184.
Sedláček, J., R. Knutti, O. Martius, and U. Beyerle, 2011: Impact of a reduced Arctic
sea ice cover on ocean and atmospheric properties. J. Clim., 25, 307–319.
Seidel, D., and W. Randel, 2007: Recent widening of the tropical belt: Evidence from
tropopause observations. J. Geophys. Res., 112, D20113.
Seidel, D. J., Q. Fu, W. J. Randel, and T. J. Reichler, 2008: Widening of the tropical belt
in a changing climate. Nature Geosci., 1, 21–24.
Sen Gupta, A., A. Santoso, A. Taschetto, C. Ummenhofer, J. Trevena, and M. England,
2009: Projected changes to the Southern Hemisphere ocean and sea ice in the
IPCC AR4 climate models. J. Clim., 22, 3047–3078.
Seneviratne, S. I., D. Lüthi, M. Litschi, and C. Schär, 2006: Land-atmosphere coupling
and climate change in Europe. Nature, 443, 205–209.
Seneviratne, S. I., et al., 2010: Investigating soil moisture-climate interactions in a
changing climate: A review. Earth Sci. Rev., 99, 125–161.
Seneviratne, S. I., et al., 2012: Changes in climate extremes and their impacts on the
natural physical environment. In: Managing the Risks of Extreme Events and
Disasters to Advance Climate Change Adaptation. A Special Report of Working
Groups I and II of the Intergovernmental Panel on Climate Change (IPCC) [C. B.
Field, et al. (eds.)]. Cambridge University Press, Cambridge, United Kingdom, and
New York, NY, USA, pp. 109–230.
Senior, C. A., and J. F. B. Mitchell, 2000: The time-dependence of climate sensitivity.
Geophys. Res. Lett., 27, 2685–2688.
Serreze, M., A. Barrett, J. Stroeve, D. Kindig, and M. Holland, 2009: The emergence of
surface-based Arctic amplification. Cryosphere, 3, 11–19.
Serreze, M. C., and J. A. Francis, 2006: The Arctic amplification debate. Clim. Change,
76, 241–264.
Sexton, D., H. Grubb, K. Shine, and C. Folland, 2003: Design and analysis of climate
model experiments for the efficient estimation of anthropogenic signals. J. Clim.,
16, 1320–1336.
Sexton, D. M. H., and J. M. Murphy, 2012: Multivariate probabilistic projections using
imperfect climate models. Part II: Robustness of methodological choices and
consequences for climate sensitivity. Clim. Dyn., 2543–2558.
Sexton, D. M. H., J. M. Murphy, M. Collins, and M. J. Webb, 2012: Multivariate
probabilistic projections using imperfect climate models. Part I: Outline of
methodology. Clim. Dyn., 2513–2542.
Shepherd, T. G., and C. McLandress, 2011: A robust mechanism for strengthening
of the Brewer-Dobson circulation in response to climate change: Critical-layer
control of subtropical wave breaking. J. Atmos. Sci., 68, 784–797.
Sherwood, S. C., 2010: Direct versus indirect effects of tropospheric humidity
changes on the hydrologic cycle. Environ. Res. Lett., 5, 025206.
1133
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Sherwood, S. C., and M. Huber, 2010: An adaptability limit to climate change due to
heat stress. Proc. Natl. Acad. Sci. U.S.A., 107, 9552–9555.
Sherwood, S. C., W. Ingram, Y. Tsushima, M. Satoh, M. Roberts, P. L. Vidale, and P. A.
O’Gorman, 2010: Relative humidity changes in a warmer climate. J. Geophys.
Res., 115, D09104.
Shindell, D., et al., 2012: Simultaneously mitigating near-term climate change and
improving human health and food security. Science, 335, 183–189.
Shindell, D. T., et al., 2006: Simulations of preindustrial, present-day, and 2100
conditions in the NASA GISS composition and climate model G-PUCCINI. Atmos.
Chem. Phys., 6, 4427–4459.
Shindell, D. T., et al., 2013a: Interactive ozone and methane chemistry in GISS-E2
historical and future climate simulations. Atmos. Chem. Phys., 13, 2653–2689.
Shindell, D. T., et al., 2013b: Radiative forcing in the ACCMIP historical and future
climate simulations. Atmos. Chem. Phys., 13, 2939–2974.
Shine, K. P., J. Cook, E. J. Highwood, and M. M. Joshi, 2003: An alternative to radiative
forcing for estimating the relative importance of climate change mechanisms.
Geophys. Res. Lett., 30, 2047.
Shiogama, H., S. Emori, K. Takahashi, T. Nagashima, T. Ogura, T. Nozawa, and T.
Takemura, 2010a: Emission scenario dependency of precipitation on global
warming in the MIROC3.2 model. J. Clim., 23, 2404–2417.
Shiogama, H., et al., 2010b: Emission scenario dependencies in climate change
assessments of the hydrological cycle. Clim. Change, 99, 321–329.
Shkolnik, I., E. Nadyozhina, T. Pavlova, E. Molkentin, and A. Semioshina, 2010: Snow
cover and permafrost evolution in Siberia as simulated by the MGO regional
climate model in the 20th and 21st centuries. Environ. Res. Lett., 5, 015005.
Shongwe, M. E., G. J. van Oldenborgh, B. van den Hurk, and M. van Aalst, 2011:
Projected changes in mean and extreme precipitation in Africa under global
warming. Part II: East Africa. J. Clim., 24, 3718–3733.
Siegenthaler, U., and H. Oeschger, 1984: Transient temperature changes due to
increasing CO
2
using simple models. Ann. Glaciol., 5, 153–159.
Sigmond, M., P. C. Siegmund, E. Manzini, and H. Kelder, 2004: A simulation of the
separate climate effects of middle-atmosphere and tropospheric CO
2
doubling.
J. Clim., 17, 2352–2367.
Sillmann, J., and E. Roeckner, 2008: Indices for extreme events in projections of
anthropogenic climate change. Clim. Change, 86, 83–104.
Sillmann, J., and M. Croci-Maspoli, 2009: Present and future atmospheric blocking
and its impact on European mean and extreme climate. Geophys. Res. Lett., 36,
L10702.
Sillmann, J., V. V. Kharin, F. W. Zwiers, X. Zhang, and D. Bronaugh, 2013: Climate
extremes indices in the CMIP5 multimodel ensemble: Part 2. Future climate
projections. J. Geophys. Res., 118, 2473–2493.
Simmons, A. J., K. M. Willett, P. D. Jones, P. W. Thorne, and D. P. Dee, 2010: Low-
frequency variations in surface atmospheric humidity, temperature, and
precipitation: Inferences from reanalyses and monthly gridded observational
data sets. J. Geophys. Res., 115, D01110.
Simpkins, G. R., and A. Y. Karpechko, 2012: Sensitivity of the southern annular mode
to greenhouse gas emission scenarios. Clim. Dyn., 38, 563–572.
Slater, A. G., and D. M. Lawrence, 2013: Diagnosing present and future permafrost
from climate models. J. Clim., doi:10.1175/JCLI-D-12-00341.1.
Smeets, E. M. W., L. F. Bouwmanw, E. Stehfest, D. P. van Vuuren, and A. Posthuma,
2009: Contribution of N
2
O to the greenhouse gas balance of first-generation
biofuels. Global Change Biol., 15, 1–23.
Smith, L., Y. Sheng, G. MacDonald, and L. Hinzman, 2005: Disappearing Arctic lakes.
Science, 308, 1429–1429.
Smith, L., et al., 2004: Siberian peatlands a net carbon sink and global methane
source since the early Holocene. Science, 303, 353–356.
Smith, R. L., C. Tebaldi, D. Nychka, and L. O. Mearns, 2009: Bayesian modeling of
uncertainty in ensembles of climate models. J. Am. Stat. Assoc., 104, 97–116.
Smith, S. J., J. van Aardenne, Z. Klimont, R. J. Andres, A. Volke, and S. Delgado Arias,
2011: Anthropogenic sulfur dioxide emissions: 1850–2005. Atmos. Chem. Phys.,
11, 1101–1116.
Smith, S. M., J. A. Lowe, N. H. A. Bowerman, L. K. Gohar, C. Huntingford, and M.
R. Allen, 2012: Equivalence of greenhouse-gas emissions for peak temperature
limits. Nature Clim. Change, 2, 535–538.
Sobel, A. H., and S. J. Camargo, 2011: Projected future seasonal changes in tropical
summer climate. J. Clim., 24, 473–487.
Soden, B., I. Held, R. Colman, K. Shell, J. Kiehl, and C. Shields, 2008: Quantifying
climate feedbacks using radiative kernels. J. Clim., 21, 3504–3520.
Soden, B. J., and I. M. Held, 2006: An assessment of climate feedbacks in coupled
ocean-atmosphere models. J. Clim., 19, 3354–3360.
Soden, B. J., and G. A. Vecchi, 2011: The vertical distribution of cloud feedback in
coupled ocean-atmosphere models. Geophys. Res. Lett., 38, L12704.
Sohn, B. J., and S.-C. Park, 2010: Strengthened tropical circulations in past three
decades inferred from water vapor transport. J. Geophys. Res., 115, D15112.
Sokolov, A. P., et al., 2009: Probabilistic forecast for twenty-first-century climate
based on uncertainties in emissions (without policy) and climate parameters. J.
Clim., 23, 2230–2231.
Solgaard, A. M., and P. L. Langen, 2012: Multistability of the Greenland ice sheet and
the effects of an adaptive mass balance formulation. Clim. Dyn., 39, 1599–1612.
Solomon, S., G. Plattner, R. Knutti, and P. Friedlingstein, 2009: Irreversible climate
change due to carbon dioxide emissions. Proc. Natl. Acad. Sci. U.S.A., 106,
1704–1709.
Solomon, S., J. Daniel, T. Sanford, D. Murphy, G. Plattner, R. Knutti, and P. Friedlingstein,
2010: Persistence of climate changes due to a range of greenhouse gases. Proc.
Natl. Acad. Sci. U.S.A., 107, 18354–18359.
Solomon, S., et al., 2007: Technical Summary. In: Climate Change 2007: The Physical
Science Basis. Contribution of Working Group I to the Fourth Assessment Report
of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M.
Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor and H. L. Miller (eds.)]
Cambridge University Press, Cambridge, United Kingdom and New York, NY,
USA, pp. 19–92.
Son, S. W., et al., 2010: Impact of stratospheric ozone on Southern Hemisphere
circulation change: A multimodel assessment. J. Geophys. Res., 115, D00M07.
Sorensson, A., C. Menendez, R. Ruscica, P. Alexander, P. Samuelsson, and U.
Willen, 2010: Projected precipitation changes in South America: A dynamical
downscaling within CLARIS. Meteorol. Z., 19, 347–355.
Spence, P., O. A. Saenko, C. O. Dufour, J. Le Sommer, and M. H. England, 2012:
Mechanisms maintaining Southern Ocean meridional heat transport under
projected wind forcing. J. Phys. Oceanogr., 42, 1923–1931.
St. Clair, S., J. Hillier, and P. Smith, 2008: Estimating the pre-harvest greenhouse gas
costs of energy crop production. Biomass Bioenerg., 32, 442–452.
Stachnik, J. P., and C. Schumacher, 2011: A comparison of the Hadley circulation in
modern reanalyses. J. Geophys. Res., 116, D22102.
Stephenson, D. B., M. Collins, J. C. Rougier, and R. E. Chandler, 2012: Statistical
problems in the probabilistic prediction of climate change. Environmetrics, 23,
364–372.
Sterl, A., et al., 2008: When can we expect extremely high surface temperatures?
Geophys. Res. Lett., 35, L14703.
Stott, P., G. Jones, and J. Mitchell, 2003: Do models underestimate the solar
contribution to recent climate change? J. Clim., 16, 4079–4093.
Stott, P., P. Good, G. A. Jones, N. Gillett, and E. Hawkins, 2013: The upper end of
climate model temperature projections is inconsistent with past warming.
Environ. Res. Lett., 8, 014024.
Stouffer, R., 2004: Time scales of climate response. J. Clim., 17, 209–217.
Stouffer, R. J., and S. Manabe, 1999: Response of a coupled ocean-atmosphere model
to increasing atmospheric carbon dioxide: Sensitivity to the rate of increase. J.
Clim., 12, 2224–2237.
Stowasser, M., H. Annamalai, and J. Hafner, 2009: Response of the South Asian
summer monsoon to global warming: Mean and synoptic systems. J. Clim., 22,
1014–1036.
Stroeve, J., M. Holland, W. Meier, T. Scambos, and M. Serreze, 2007: Arctic sea ice
decline: Faster than forecast. Geophys. Res. Lett., 34, L09501.
Stroeve, J. C., V. Kattsov, A. Barrett, M. Serreze, T. Pavlova, M. Holland, and W. N.
Meier, 2012: Trends in Arctic sea ice extent from CMIP5, CMIP3 and observations.
Geophys. Res. Lett., 39, L16502.
Stuber, N., M. Ponater, and R. Sausen, 2005: Why radiative forcing might fail as a
predictor of climate change. Clim. Dyn., 24, 497–510.
Sudo, K., M. Takahashi, and H. Akimoto, 2003: Future changes in stratosphere-
troposphere exchange and their impacts on future tropospheric ozone
simulations. Geophys. Res. Lett., 30, 2256.
Sugiyama, M., H. Shiogama, and S. Emori, 2010: Precipitation extreme changes
exceeding moisture content increases in MIROC and IPCC climate models. Proc.
Natl. Acad. Sci. U.S.A., 107, 571–575.
Sun, Y., S. Solomon, A. Dai, and R. W. Portmann, 2007: How often will it rain? J. Clim.,
20, 4801–4818.
1134
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Sutton, R. T., B. W. Dong, and J. M. Gregory, 2007: Land/sea warming ratio in response
to climate change: IPCC AR4 model results and comparison with observations.
Geophys. Res. Lett., 34, L02701.
Swann, A. L., I. Y. Fung, S. Levis, G. B. Bonan, and S. C. Doney, 2010: Changes in Arctic
vegetation amplify high-latitude warming through the greenhouse effect. Proc.
Natl. Acad. Sci. U.S.A., 107, 1295–1300.
Swart, N. C., and J. C. Fyfe, 2012: Observed and simulated changes in the Southern
Hemisphere surface westerly wind-stress. Geophys. Res. Lett., 39, L16711.
Swingedouw, D., P. Braconnot, P. Delecluse, E. Guilyardi, and O. Marti, 2007:
Quantifying the AMOC feedbacks during a 2xCO
2
stabilization experiment with
land-ice melting. Clim. Dyn., 29, 521–534.
Swingedouw, D., T. Fichefet, P. Huybrechts, H. Goosse, E. Driesschaert, and M. Loutre,
2008: Antarctic ice-sheet melting provides negative feedbacks on future climate
warming. Geophys. Res. Lett., 35, L17705.
Szopa, S., et al., 2013: Aerosol and ozone changes as forcing for climate evolution
between 1850 and 2100. Clim. Dyn., 40, 2223–2250.
Takahashi, K., 2009a: Radiative constraints on the hydrological cycle in an idealized
radiative-convective equilibrium model. J. Atmos. Sci., 66, 77–91.
Takahashi, K., 2009b: The global hydrological cycle and atmospheric shortwave
absorption in climate models under CO
2
forcing. J. Clim., 22, 5667–5675.
Tanaka, K., and T. Raddatz, 2011: Correlation between climate sensitivity and aerosol
forcing and its implication for the “climate trap”. Clim. Change, 109, 815–825.
Tarnocai, C., J. Canadell, E. Schuur, P. Kuhry, G. Mazhitova, and S. Zimov, 2009: Soil
organic carbon pools in the northern circumpolar permafrost region. Global
Biogeochem. Cycles, 23, GB2023.
Taylor, K. E., R. J. Stouffer, and G. A. Meehl, 2012: A summary of the CMIP5 experiment
design. Bull. Am. Meteorol. Soc., 93, 485–498.
Tebaldi, C., and R. Knutti, 2007: The use of the multi-model ensemble in probabilistic
climate projections. Philos. Trans. R. Soc. A, 365, 2053–2075.
Tebaldi, C., and D. B. Lobell, 2008: Towards probabilistic projections of climate
change impacts on global crop yields. Geophys. Res. Lett., 35, L08705.
Tebaldi, C., and B. Sanso, 2009: Joint projections of temperature and precipitation
change from multiple climate models: A hierarchical Bayesian approach. J. R.
Stat. Soc. A, 172, 83–106.
Tebaldi, C., J. M. Arblaster, and R. Knutti, 2011: Mapping model agreement on future
climate projections. Geophys. Res. Lett., 38, L23701.
Tebaldi, C., K. Hayhoe, J. M. Arblaster, and G. A. Meehl, 2006: Going to the extremes.
Clim. Change, 79, 185–211.
Terray, L., L. Corre, S. Cravatte, T. Delcroix, G. Reverdin, and A. Ribes, 2012: Near-
surface salinity as nature’s rain gauge to detect human influence on the tropical
water cycle. J. Clim., 25, 958–977.
Thorne, P., 2008: Arctic tropospheric warming amplification? Nature, 455, E1–E2.
Tietsche, S., D. Notz, J. H. Jungclaus, and J. Marotzke, 2011: Recovery mechanisms of
Arctic summer sea ice. Geophys. Res. Lett., 38, L02707.
Tjiputra, J. F., et al., 2013: Evaluation of the carbon cycle components in the
Norwegian Earth System Model (NorESM). Geosci. Model Dev., 6, 301–325.
Tokinaga, H., S.-P. Xie, C. Deser, Y. Kosaka, and Y. M. Okumura, 2012: Slowdown
of the Walker circulation driven by tropical Indo-Pacific warming. Nature, 491,
439–443.
Trapp, R. J., N. S. Diffenbaugh, and A. Gluhovsky, 2009: Transient response of severe
thunderstorm forcing to elevated greenhouse gas concentrations. Geophys. Res.
Lett., 36, L01703.
Trapp, R. J., N. S. Diffenbaugh, H. E. Brooks, M. E. Baldwin, E. D. Robinson, and J. S. Pal,
2007: Changes in severe thunderstorm environment frequency during the 21st
century caused by anthropogenically enhanced global radiative forcing. Proc.
Natl. Acad. Sci. U.S.A., 104, 19719–19723.
Trenberth, K. E., and D. J. Shea, 2005: Relationships between precipitation and
surface temperature. Geophys. Res. Lett., 32, L14703.
Trenberth, K. E., and J. T. Fasullo, 2009: Global warming due to increasing absorbed
solar radiation. Geophys. Res. Lett., 36, L07706.
Trenberth, K. E., and J. T. Fasullo, 2010: Simulation of present-day and twenty-first-
century energy budgets of the Southern Oceans. J. Clim., 23, 440–454.
Turner, J., T. J. Bracegirdle, T. Phillips, G. J. Marshall, and J. S. Hosking, 2013: An initial
assessment of Antarctic sea ice extent in the CMIP5 models. J. Clim., 26, 1473–
1484.
Ueda, H., A. Iwai, K. Kuwako, and M. Hori, 2006: Impact of anthropogenic forcing on
the Asian summer monsoon as simulated by eight GCMs. Geophys. Res. Lett.,
33, L06703.
Ulbrich, U., G. C. Leckebusch, and J. G. Pinto, 2009: Extra-tropical cyclones in the
present and future climate: A review. Theor. Appl. Climatol., 96, 117–131.
Ulbrich, U., et al., 2013: Are greenhouse gas signals of Northern Hemisphere winter
extra-tropical cyclone activity dependent on the identification and tracking
algorithm? Meteorol. Z., 22, 61–68.
UNEP, 2010: The emissions gap report: Are the Copenhagen Accord pledges sufficient
to limit global warming to 2°C or 1.5°C? , 55 pp.
Utsumi, N., S. Seto, S. Kanae, E. E. Maeda, and T. Oki, 2011: Does higher surface
temperature intensify extreme precipitation? Geophys. Res. Lett., 38, L16708.
Vaks, A., et al., 2013: Speleothems reveal 500,000-year history of Siberian
permafrost. Science, 340, 183–186.
Van Klooster, S. L., and P. J. Roebber, 2009: Surface-based convective potential in
the contiguous United States in a business-as-usual future climate. J. Clim., 22,
3317–3330.
van Vuuren, D. P., et al., 2011: RCP3–PD: Exploring the possibilities to limit global
mean temperature change to less than 2°C. Clim. Change, 109, 95–116.
Vavrus, S., M. Holland, and D. Bailey, 2011: Changes in Arctic clouds during intervals
of rapid sea ice loss. Clim. Dyn., 36, 1475–1489.
Vavrus, S. J., M. M. Holland, A. Jahn, D. A. Bailey, and B. A. Blazey, 2012: Twenty-first-
century Arctic climate change in CCSM4. J. Clim., 25, 2696–2710.
Vecchi, G. A., and B. J. Soden, 2007: Global warming and the weakening of the
tropical circulation. Bull. Am. Meteorol. Soc., 88, 1529–1530.
Vecchi, G. A., B. J. Soden, A. T. Wittenberg, I. M. Held, A. Leetmaa, and M. J. Harrison,
2006: Weakening of tropical Pacific atmospheric circulation due to anthropogenic
forcing. Nature, 441, 73–76.
Vial, J., J.-L. Dufresne, and S. Bony, 2013: On the interpretation of inter-model spread
in CMIP5 climate sensitivity estimates. Clim. Dyn., doi:10.1007/s00382-013-
1725-9.
Vidale, P. L., D. Lüthi, R. Wegmann, and C. Schär, 2007: European summer climate
variability in a heterogeneous multi-model ensemble. Clim. Change, 81, 209–
232.
Voldoire, A., et al., 2013: The CNRM-CM5.1 global climate model: Description and
basic evaluation. Clim. Dyn., 40, 2091–2121.
Voss, R., and U. Mikolajewicz, 2001: Long-term climate changes due to increased
CO
2
concentration in the coupled atmosphere-ocean general circulation model
ECHAM3/LSG. Clim. Dyn., 17, 45–60.
Wadhams, P., 2012: Arctic ice cover, ice thickness and tipping points. Ambio, 41,
23–33.
Walker, R., et al., 2009: Protecting the Amazon with protected areas. Proc. Natl. Acad.
Sci. U.S.A., 106, 10582–10586.
Wang, M., and J. Overland, 2009: A sea ice free summer Arctic within 30 years?
Geophys. Res. Lett., 36, L07502.
Wang, M., and J. E. Overland, 2012: A sea ice free summer Arctic within 30 years: An
update from CMIP5 models. Geophys. Res. Lett., 39, L18501.
Wania, R., I. Ross, and I. Prentice, 2009: Integrating peatlands and permafrost into
a dynamic global vegetation model: 2. Evaluation and sensitivity of vegetation
and carbon cycle processes. Global Biogeochem. Cycles, 23, GB3015.
Washington, W., et al., 2009: How much climate change can be avoided by
mitigation? Geophys. Res. Lett., 36, L08703.
Watanabe, S., et al., 2011: MIROC-ESM 2010: Model description and basic results of
CMIP5-20c3m experiments. Geosci. Model Dev., 4, 845–872.
Watterson, I. G., 2008: Calculation of probability density functions for temperature
and precipitation change under global warming. J. Geophys. Res., 113, D12106.
Watterson, I. G., 2011: Calculation of joint PDFs for climate change with properties
matching recent Australian projections. Aust. Meteorol. Oceanogr. J., 61, 211–
219.
Watterson, I. G., and P. H. Whetton, 2011a: Joint PDFs for Australian climate in
future decades and an idealized application to wheat crop yield. Aust. Meteorol.
Oceanogr. J., 61, 221–230.
Watterson, I. G., and P. H. Whetton, 2011b: Distributions of decadal means of
temperature and precipitation change under global warming. J. Geophys. Res.,
116, D07101.
Watterson, I. G., J. L. McGregor, and K. C. Nguyen, 2008: Changes in extreme
temperatures of Australasian summer simulated by CCAM under global
warming, and the roles of winds and land-sea contrasts. Aust. Meteorol. Mag.,
57, 195–212.
WBGU, 2009: Solving the Climate Dilemma: The Budget Approach. German Advisory
Council on Global Change, Berlin, 59 pp.
1135
Long-term Climate Change: Projections, Commitments and Irreversibility Chapter 12
12
Weaver, A., K. Zickfeld, A. Montenegro, and M. Eby, 2007: Long term climate
implications of 2050 emission reduction targets. Geophys. Res. Lett., 34, L19703.
Weaver, A. J., et al., 2012: Stability of the Atlantic meridional overturning circulation:
A model intercomparison. Geophys. Res. Lett., 39, L20709.
Webb, M., et al., 2006: On the contribution of local feedback mechanisms to the
range of climate sensitivity in two GCM ensembles. Clim. Dyn., 27, 17–38.
Webb, M. J., F. H. Lambert, and J. M. Gregory, 2013: Origins of differences in climate
sensitivity, forcing and feedback in climate models. Clim. Dyn., 40, 677–707.
Weber, S., et al., 2007: The modern and glacial overturning circulation in the Atlantic
Ocean in PMIP coupled model simulations. Clim. Past, 3, 51–64.
Weertman, J., 1974: Stability of the junction of an ice sheet and an ice shelf. J.
Glaciol., 13, 3–11.
Wehner, M., D. Easterling, J. Lawrimore, R. Heim, R. Vose, and B. Santer, 2011:
Projections of future drought in the continental United States and Mexico. J.
Hydrometeorol., 12, 1359–1377.
Weigel, A., R. Knutti, M. Liniger, and C. Appenzeller, 2010: Risks of model weighting
in multimodel climate projections. J. Clim., 23, 4175–4191.
Wentz, F., L. Ricciardulli, K. Hilburn, and C. Mears, 2007: How much more rain will
global warming bring? Science, 317, 233–235.
Wetherald, R., and S. Manabe, 1988: Cloud feedback processes in a General-
Circulation Model. J. Atmos. Sci., 45, 1397–1415.
Wetherald, R. T., R. J. Stouffer, and K. W. Dixon, 2001: Committed warming and its
implications for climate change. Geophys. Res. Lett., 28, 1535–1538.
Wigley, T. M. L., 2005: The climate change commitment. Science, 307, 1766–1769.
Wilcox, L. J., A. J. Charlton-Perez, and L. J. Gray, 2012: Trends in Austral jet position in
ensembles of high- and low-top CMIP5 models. J. Geophys. Res., 117, D13115.
Willett, K., and S. Sherwood, 2012: Exceedance of heat index thresholds for 15
regions under a warming climate using the wet-bulb globe temperature. Int. J.
Climatol., 32, 161–177.
Williams, J. W., S. T. Jackson, and J. E. Kutzbach, 2007: Projected distributions of
novel and disappearing climates by 2100 AD. Proc. Natl. Acad. Sci. U.S.A., 104,
5738–5742.
Williams, K. D., W. J. Ingram, and J. M. Gregory, 2008: Time variation of effective
climate sensitivity in GCMs. J. Clim., 21, 5076–5090.
Winton, M., 2006a: Amplified Arctic climate change: What does surface albedo
feedback have to do with it? Geophys. Res. Lett., 33, L03701.
Winton, M., 2006b: Does the Arctic sea ice have a tipping point? Geophys. Res. Lett.,
33, L23504.
Winton, M., 2008: Sea ice-albedo feedback and nonlinear Arctic climate change. In:
Arctic Sea Ice Decline: Observations, Projections, Mechanisms, and Implications
[E. T. DeWeaver, C. M. Bitz and L. B. Tremblay (eds.)]. American Geophysical
Union, Washington, DC, pp. 111–131.
Winton, M., 2011: Do climate models underestimate the sensitivity of Northern
Hemisphere sea ice cover? J. Clim., 24, 3924–3934.
WMO, 2007: Scientific assessment of ozone depletion. In: 2006, Global Ozone
Research and Monitoring Project. World Meteorological Organization, Geneva,
Switzerland, 572 pp.
Wood, R., A. Keen, J. Mitchell, and J. Gregory, 1999: Changing spatial structure of
the thermohaline circulation in response to atmospheric CO
2
forcing in a climate
model. Nature, 399, 572–575.
Woollings, T., 2008: Vertical structure of anthropogenic zonal-mean atmospheric
circulation change. Geophys. Res. Lett., 35, L19702.
Woollings, T., and M. Blackburn, 2012: The North Atlantic jet stream under climate
change and its relation to the NAO and EA patterns. J. Clim., 25, 886–902.
Woollings, T., J. M. Gregory, J. G. Pinto, M. Reyers, and D. J. Brayshaw, 2012: Response
of the North Atlantic storm track to climate change shaped by ocean-atmosphere
coupling. Nature Geosci., 5, 313–317.
Wu, P., R. Wood, J. Ridley, and J. Lowe, 2010: Temporary acceleration of the
hydrological cycle in response to a CO
2
rampdown. Geophys. Res. Lett., 37,
L12705.
Wu, P., L. Jackson, A. Pardaens, and N. Schaller, 2011a: Extended warming of
the northern high latitudes due to an overshoot of the Atlantic meridional
overturning circulation. Geophys. Res. Lett., 38, L24704.
Wu, T., et al., 2013: Global carbon budgets simulated by the Beijing Climate Center
Climate System Model for the last century. J. Geophys. Res., doi:10.1002/
jgrd.50320.
Wu, Y., M. Ting, R. Seager, H.-P. Huang, and M. A. Cane, 2011b: Changes in storm
tracks and energy transports in a warmer climate simulated by the GFDL CM2.1
model. Clim. Dyn., 37, 53–72.
Wyant, M. C., et al., 2006: A comparison of low-latitude cloud properties and their
response to climate change in three AGCMs sorted into regimes using mid-
tropospheric vertical velocity. Clim. Dyn., 27, 261–279.
Xie, P., and G. Vallis, 2012: The passive and active nature of ocean heat uptake in
idealized climate change experiments. Clim. Dyn., 38, 667–684.
Xie, S. P., C. Deser, G. A. Vecchi, J. Ma, H. Y. Teng, and A. T. Wittenberg, 2010: Global
warming pattern formation: Sea surface temperature and rainfall. J. Clim., 23,
966–986.
Xin, X., L. Zhang, J. Zhang, T. Wu, and Y. Fang, 2013a: Climate change projections over
East Asia with BCC_CSM1.1 climate model under RCP scenarios. J. Meteorol.
Soc. Jpn., 4, 413-429.
Xin, X., T. Wu, J. Li, Z. Wang, W. Li, and F. Wu, 2013b: How well does BCC_CSM1.1
reproduce the 20th century climate change in China? Atmos. Ocean. Sci. Lett.,
6, 21−26.
Yang, F. L., A. Kumar, M. E. Schlesinger, and W. Q. Wang, 2003: Intensity of hydrological
cycles in warmer climates. J. Clim., 16, 2419–2423.
Yin, J., J. Overpeck, S. Griffies, A. Hu, J. Russell, and R. Stouffer, 2011: Different
magnitudes of projected subsurface ocean warming around Greenland and
Antarctica. Nature Geosci., 4, 524–528.
Yokohata, T., M. Webb, M. Collins, K. Williams, M. Yoshimori, J. Hargreaves, and J.
Annan, 2010: Structural similarities and differences in climate responses to CO
2
increase between two perturbed physics ensembles. J. Clim., 23, 1392–1410.
Yokohata, T., J. D. Annan, M. Collins, C. S. Jackson, M. Tobis, M. Webb, and J. C.
Hargreaves, 2012: Reliability of multi-model and structurally different single-
model ensembles. Clim. Dyn., 39, 599–616.
Young, P. J., K. H. Rosenlof, S. Solomon, S. C. Sherwood, Q. Fu, and J.-F. Lamarque,
2012: Changes in stratospheric temperatures and their implications for changes
in the Brewer Dobson circulation, 1979–2005. J. Clim., 25, 1759–1772.
Yukimoto, S., et al., 2012: A new global climate model of the Meteorological
Research Institute: MRI-CGCM3–Model description and basic performance. J.
Meteorol. Soc. Jpn., 90A, 23–64.
Zappa, G., L. C. Shaffrey, K. I. Hodges, P. G. Sansom, and D. B. Stephenson, 2013: A
multi-model assessment of future projections of North Atlantic and European
extratropical cyclones in the CMIP5 climate models. J. Clim., doi:10.1175/JCLI-
D-12-00573.1.
Zelazowski, P., Y. Malhi, C. Huntingford, S. Sitch, and J. Fisher, 2011: Changes in the
potential distribution of humid tropical forests on a warmer planet. Philos. Trans.
R. Soc. A, 369, 137–160.
Zelinka, M., and D. Hartmann, 2010: Why is longwave cloud feedback positive? J.
Geophys. Res., 115, D16117.
Zelinka, M., S. Klein, and D. Hartmann, 2012: Computing and partitioning cloud
feedbacks using Cloud property histograms. Part II: Attribution to changes in
cloud amount, altitude, and optical depth. J. Clim., 25, 3736–3754.
Zhang, M., and H. Song, 2006: Evidence of deceleration of atmospheric vertical
overturning circulation over the tropical Pacific. Geophys. Res. Lett., 33, L12701.
Zhang, M. H., and C. Bretherton, 2008: Mechanisms of low cloud-climate feedback
in idealized single-column simulations with the Community Atmospheric Model,
version 3 (CAM3). J. Clim., 21, 4859–4878.
Zhang, R., 2010a: Northward intensification of anthropogenically forced changes
in the Atlantic meridional overturning circulation (AMOC). Geophys. Res. Lett.,
37, L24603.
Zhang, T., 2005: Influence of the seasonal snow cover on the ground thermal regime:
An overview. Rev. Geophys., 43, RG4002.
Zhang, T., J. A. Heginbottom, R. G. Barry, and J. Brown, 2000: Further statistics on the
distribution of permafrost and ground ice in the Northern Hemisphere 1. Polar
Geogr., 24, 126–131.
Zhang, X., 2010b: Sensitivity of Arctic summer sea ice coverage to global warming
forcing: Towards reducing uncertainty in arctic climate change projections. Tellus
A, 62, 220–227.
Zhang, X., and J. Walsh, 2006: Toward a seasonally ice-covered Arctic Ocean:
Scenarios from the IPCC AR4 model simulations. J. Clim., 19, 1730–1747.
Zhang, X. B., et al., 2007: Detection of human influence on twentieth-century
precipitation trends. Nature, 448, 461–U464.
Zhou, L. M., R. E. Dickinson, P. Dirmeyer, A. Dai, and S. K. Min, 2009: Spatiotemporal
patterns of changes in maximum and minimum temperatures in multi-model
simulations. Geophys. Res. Lett., 36, L02702.
1136
Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility
12
Zhuang, Q., et al., 2004: Methane fluxes between terrestrial ecosystems and the
atmosphere at northern high latitudes during the past century: A retrospective
analysis with a process-based biogeochemistry model. Global Biogeochem.
Cycles, 18, GB3010.
Zickfeld, K., V. K. Arora, and N. P. Gillett, 2012: Is the climate response to CO
2
emissions path dependent? Geophys. Res. Lett., 39, L05703.
Zickfeld, K., B. Knopf, V. Petoukhov, and H. Schellnhuber, 2005: Is the Indian summer
monsoon stable against global change? Geophys. Res. Lett., 32, L15707.
Zickfeld, K., M. Eby, H. Matthews, and A. Weaver, 2009: Setting cumulative emissions
targets to reduce the risk of dangerous climate change. Proc. Natl. Acad. Sci.
U.S.A., 106, 16129–16134.
Zickfeld, K., M. Morgan, D. Frame, and D. Keith, 2010: Expert judgments about
transient climate response to alternative future trajectories of radiative forcing.
Proc. Natl. Acad. Sci. U.S.A., 107, 12451–12456.
Zickfeld, K., et al., 2013: Long-term climate change commitment and reversibility: An
EMIC intercomparison. J. Clim., doi:10.1175/JCLI-D-12-00584.1.
Zimov, S., S. Davydov, G. Zimova, A. Davydova, E. Schuur, K. Dutta, and F. Chapin,
2006: Permafrost carbon: Stock and decomposability of a globally significant
carbon pool. Geophys. Res. Lett., 33, L20502.
Zunz, V., H. Goosse, and F. Massonnet, 2013: How does internal variability influence
the ability of CMIP5 models to reproduce the recent trend in Southern Ocean sea
ice extent? Cryosphere, 7, 451–468.