Reframing the climate change challenge in light of post-2000 emission trends

Abstract

The 2007 Bali conference heard repeated calls for reductions in global greenhouse gas emissions of 50 per cent by 2050 to avoid exceeding the 2 degrees C threshold. While such endpoint targets dominate the policy agenda, they do not, in isolation, have a scientific basis and are likely to lead to dangerously misguided policies. To be scientifically credible, policy must be informed by an understanding of cumulative emissions and associated emission pathways. This analysis considers the implications of the 2 degrees C threshold and a range of post-peak emission reduction rates for global emission pathways and cumulative emission budgets. The paper examines whether empirical estimates of greenhouse gas emissions between 2000 and 2008, a period typically modelled within scenario studies, combined with short-term extrapolations of current emissions trends, significantly constrains the 2000-2100 emission pathways. The paper concludes that it is increasingly unlikely any global agreement will deliver the radical reversal in emission trends required for stabilization at 450 ppmv carbon dioxide equivalent (CO2e). Similarly, the current framing of climate change cannot be reconciled with the rates of mitigation necessary to stabilize at 550 ppmv CO2e and even an optimistic interpretation suggests stabilization much below 650 ppmv CO2e is improbable.

Full text

Reframing the climate change challenge in
light of post-2000 emission trends
BYKEVIN ANDERSON AND ALICE BOWS*
Tyndall Centre for Climate Change Research, Mechanical, Civil and Aerospace
Engineering, University of Manchester, PO Box 88, Manchester M60 1QD, UK
The 2007 Bali conference heard repeated calls for reductions in global greenhouse gas
emissions of 50 per cent by 2050 to avoid exceeding the 28C threshold. While such endpoint
targets dominate the policy agenda, they do not, in isolation, have a scientific basis and are
likely to lead to dangerously misguided policies. To be scientifically credible, policy must be
informed by an understanding of cumulative emissions and associated emission pathways.
This analysis considers the implications of the 28C threshold and a range of post-peak
emission reduction rates for global emission pathways and cumulative emission budgets.
The paper examines whether empirical estimates of greenhouse gas emissions between 2000
and 2008, a period typically modelled within scenario studies, combined with short-term
extrapolations of current emissions trends, significantly constrains the 2000–2100 emission
pathways. The paper concludes that it is increasingly unlikely any global agreement will
deliver the radical reversal in emission trends required for stabilization at 450 ppmv carbon
dioxide equivalent (CO
2
e). Similarly, the current framing of climate change cannot be
reconciled with the rates of mitigation necessary to stabilize at 550 ppmv CO
2
e and even an
optimistic interpretation suggests stabilization much below 650 ppmv CO
2
e is improbable.
Keywords: emission scenarios; cumulative emissions; climate policy; energy;
emission trends
1. Introduction
In the absence of global agreement on a metric for delineating dangerous from
acceptable climate change, 28C has, almost by default, emerged as the principal
focus of international and national policy.
1
Moreover, within the scientific
community, 28C has come to provide a benchmark temperature against which to
consider atmospheric concentrations of greenhouse gases and emission reduction
profiles. While it is legitimate to question whether temperature is an appropriate
metric for representing climate change and, if it is, whether 28C is the appro-
priate temperature (Tol 2007), this is not the purpose of this paper. Instead, the
paper begins by considering the implications of the 28C threshold for global
emission pathways, before proceeding to consider the implications of different
emission pathways on stabilization concentrations and associated temperatures.
Phil. Trans. R. Soc. A
doi:10.1098/rsta.2008.0138
Published online
One contribution of 12 to a Theme Issue ‘Geoscale engineering to avert dangerous climate change’.
* Author for correspondence (alice.bows@manchester.ac.uk).
1
For example, in March 2007, European leaders reaffirmed their commitment to the 28C threshold
(European Commission 2007).
1This journal is q2008 The Royal Society

Although the policy realm generally focuses on the emissions profiles between
2000 and 2050, the scientific community tends to consider longer periods,
typically up to and beyond 2100. By using a range of cumulative carbon budgets
with differing degrees of carbon-cycle feedbacks, this paper assesses whether
global emissions of greenhouse gases between 2000 and 2008, combined with
short-term extrapolations of emission trends, significantly impact the 2008–2100
cumulative emission budget available, and hence emission pathways.
In brief, the paper combines current greenhouse gas emissions data (including
deforestation) with up-to-date emission trends and the latest scientific under-
standing of the relationships between emissions and concentrations to consider
three questions.
(i) Given a small set of emissions pathways from 2000 to a date where global
emissions are assumed to peak (2015, 2020 and 2025), what emission
reduction rates would be necessary to remain within the 2000–2100
cumulative emission budgets associated with atmospheric stabilization of
carbon dioxide equivalent (CO
2
e) at 450 ppmv? The accompanying
scenario set is hereafter referred to as ‘Anderson Bows 1’ (AB1 ).
(ii) Given the same pathways from 2000 to the 2020 emissions peak, what
concentrations of CO
2
e are associated with subsequent annual emission
reduction rates of 3, 5 and 7 per cent? The accompanying scenario set is
hereafter referred to as ‘Anderson Bows 2’ (AB2).
(iii) What are the implications of the findings from (i) and (ii) for the current
framing of the climate agenda more generally, and the appropriateness
of the 28C threshold as the driver of mitigation and adaptation policy
more specifically?
2. Analysis framing
(a)Correlating 2 8C with greenhouse gas concentration and carbon budgets
What constitutes an acceptable temperature increase is a political rather than a
scientific decision, though the former may be informed by science. By contrast,
the correlation between temperature, atmospheric concentration of CO
2
e and
anthropogenic cumulative emission budgets emerges, primarily, from our
scientific understanding of how the climate functions.
According to a recent synthesis of global climate models (Meinshausen 2006,
table 28.1), the 550 ppmv CO
2
e concentration, around which much policy
discussion revolves, suggests an 82 per cent mid-value probability of exceeding
28C. By contrast, to provide a 93 per cent mid-value probability of not exceeding
28C, the concentration would need to be stabilized at, or below, 350 ppmv CO
2
e, i.e.
below current levels. While Meinshausen’s analysis demonstrates the gulf between
the science and the policy of approximately 28C, the analysis within the IPCC’s
fourth assessment report (IPCC 2007a), hereafter AR4, suggests that the scale of
the challenge is even more demanding. Not only has the ‘best estimate’ of climate
sensitivity risen from 2.58C in the 1996 report (IPCC 1996, p. 39) to 38C in AR4,
but also the inclusion of carbon-cycle feedbacks has significantly reduced the
cumulative anthropogenic emissions (carbon budget) associated with particular
concentrations of CO
2
e(IPCC 2007a, topic 5, p. 6).
K. Anderson and A. Bows2
Phil. Trans. R. Soc. A

Understanding current emission trends in particular and the links between global
temperature changes and national emission budgets more generally (sometimes
referred to as the ‘correlation trail’; Anderson & Bows 2007), is essential if policy is
to be evidence based. Currently, national and international policies are dominated
by long-term reduction targets with little regard for the cumulative carbon budget
described by particular emission pathways. Within the UK, for example, while the
government acknowledges the link between temperature and concentration, the
principal focus of its policies is on reducing emissions by 60 per cent by 2050
(excluding international aviation and shipping; Bows & Anderson 2007). Closer
examination of the UK’s relatively ‘mature’ climate change policy reveals a further
inconsistency. Within many official documents 550 ppmv CO
2
eand550ppmvCO
2
are used interchangeably,
2
with the latter equating to approximately 615 ppmv
CO
2
e (extrapolated from IPCC 2007a, topic guide 5, table 5.1); the policy
repercussions of this scale of ambiguity are substantial.
Whether considering climate change from an international, national or regional
perspective, it is essential that the associated policy debate be informed by the
latest science on the ‘correlation trail’ from temperature and atmospheric
concentrations of CO
2
e through to global carbon budgets and national emission
pathways. Without such an informed debate, the scientific and policy uncertainties
that unavoidably arise are exacerbated unnecessarily and significantly.
(b)Recent emissions data and science: impact on carbon budgets
(i) Carbon-cycle feedbacks
The atmospheric concentration of CO
2
depends not only on the quantity of
emissions emitted into the atmosphere (natural and anthropogenic), but also on
land use changes and the capacity of carbon sinks within the biosphere. As the
atmospheric concentration of CO
2
increases (at least within reasonable bounds),
so there is a net increase in its take-up rate from the atmosphere by vegetation
and the ocean. However, changes in rainfall and temperature in response to
increased atmospheric greenhouse gas concentrations affect the absorptive
capacity of natural sinks (Jones et al. 2006;Canadell et al. 2007;Le Que
´re
´
et al. 2007). While the complex and interactive nature of these effects leads to
uncertainties with regard to the size of the carbon-cycle feedbacks (Cox et al.
2006), all models studied agree that a global mean temperature increase will
reduce the biosphere’s ability to store carbon emissions over the time scales
considered here (Friedlingstein et al. 2006). Consequently, pathways to
stabilizing CO
2
concentrations that include feedbacks have lower permissible
emissions than those pathways that exclude such feedbacks. According to AR4,
for example, with feedbacks included, stabilizing at 450 ppmv CO
2
e correlates
with cumulative emissions some 27 per cent lower than without feedbacks, over a
100-year period (IPCC 2007a, topic guide 5, p. 6). The impact of this latest
science on the link between emissions and temperature is of sufficient scale to
require the emission-reduction pathways associated with particular concen-
trations and hence temperatures be revisited.
2
For example, the RCEP uses CO
2
einRCEP (2000), whereas the Energy White Paper (DTI 2006)
and Climate Change Programme (DEFRA 2006) both refer to CO
2
alone.
3Reframing the climate change challenge
Phil. Trans. R. Soc. A

(ii) Latest empirical emissions data
The current suites of emission scenarios informing the international and
national climate change agenda seldom include empirical emissions data post-
2000, choosing instead to model recent emissions; both the 2006 Stern Review
(Stern 2006, p. 231) and the UK’s 2007 draft climate change bill (DEFRA 2007)
illustrate this tendency. However, recent empirical data have shown global
emissions to have risen at rates well in excess of those contained within these and
many other emissions scenarios (Raupach et al. 2007). For example, while Stern
assumes a mean annual CO
2
e emission growth between 2000 and 2006 of
approximately 0.95 per cent, the growth rate calculated from the latest empirical
data is closer to 2.4 per cent.
3
Similarly, the UK’s draft climate change bill
(DEFRA 2007) contains an emission pathway between 2000 and 2006 in which
emissions fall, while over the same period the UK Government’s emission
inventory suggests, at best, that emissions have been stable.
A further and important revision to recent emissions data relates to
deforestation. Within many scenarios, including Stern, emissions resulting from
deforestation are estimated to be in the region of 7.3 GtCO
2
in 2000. However,
recent data have suggested this to be an overestimate, with R. A. Houghton
(2006, personal communication) having recently revised his earlier figure
downward to 5.5 GtCO
2
.
4
The impact of this reduction allied to the latest
emission data reinforces the need to revisit emission pathways.
3. Scenario analysis
(a)Overview
The scenario analysis presented within this paper is for the basket of six
greenhouse gases only and relies, principally, on the scientific understanding
contained within AR4. The analysis does not take account of the following:
— the radiative forcing impacts of aerosols and non-CO
2
aviation emissions
(e.g. emissions of NO
x
in the upper troposphere, vapour trails and cirrus
formation);
5
3
CO
2
data from the Carbon Dioxide Information Analysis Centre (CDIAC) including recent data
from G. Marland (2006, personal communication); non-CO
2
greenhouse gas data from the USA
Environmental Protection Agency (EPA 2006) including the projection for 2005, and assuming
deforestation emissions in 2005 to be 5.5 GtCO
2
(1.5 GtC), with a 0.4 per cent growth in the
preceding 5 years in line with data within the Global Forest Resources Assessment (FAO 2005).
4
FAO (2005) contains rates of tropical deforestation for the 1990s revised downward from those in
the 2000 Global Forest Resources Assessment (FAO 2000; R. A. Houghton 2006, personal
communication). An earlier estimate based on high-resolution satellite data over areas identified as
‘hot spots’ of deforestation, estimated the figure at nearer 3.7 GtCO
2
(1 GtC) for 2000 (Achard
et al. 2004). It is Houghton’s more recent estimate that is used in this paper.
5
There remains considerable uncertainty as to the actual level of radiative forcing associated
with aerosols, exacerbated by their relatively short residence times in the atmosphere and
uncertainty as to future aerosol emission pathways (Cranmer et al. 2001;Andreae et al. 2005).
Similarly, there remain significant uncertainties as to the radiative forcing impact of non-CO
2
emissions from aviation, particularly contrails and linear cirrus (e.g. Stordal et al. 2004;
Mannstein & Schumann 2005).
K. Anderson and A. Bows4
Phil. Trans. R. Soc. A

— the most recent findings with respect to carbon sinks;
6
— previously underestimated emission sources;
7
and
— the implications of early emission peaks for ‘overshooting’ stabilization
concentrations and the attendant risks of additional feedbacks.
While aerosols are most commonly associated with net global (or at least
regional) cooling, the other factors outlined above are either net positive
feedbacks or, as is the case for high peak-level emissions, increase the likelihood
of net positive feedbacks. Consequently, the correlations between concentration
and mitigation outlined in this analysis are, in time, liable to prove conservative.
The scenarios are for CO
2
e emission pathways during the twenty-first century,
with empirical data used for the opening years of the century (in contrast to
modelled or ‘what if’ data). The full scenario sets (AB1 and AB2) comprise
different combinations of the following: (i) emissions of CO
2
from deforestation,
(ii) emissions of non-CO
2
greenhouse gases, and (iii) emissions of CO
2
from
energy and industrial processes.
For AB1
—Deforestation. Two low emission scenarios for the twenty-first century.
—Non-CO
2
greenhouse gases. Three scenarios peaking in 2015, 2020 and 2025
and subsequently reducing to 7.5 GtCO
2
e per year.
—Energy and process CO
2
. Three scenarios peaking in 2015, 2020 and 2025 and
subsequently reducing to maintain the total cumulative emissions for the
twenty-first century within the AR4 450 ppmv CO
2
e range (with carbon-cycle
feedbacks).
For AB2
—Deforestation. Two low emission scenarios for the twenty-first century.
—Non-CO
2
greenhouse gases. One scenario peaking in 2020 subsequently
reducing to 7.5 GtCO
2
e per year (as per AB1 with a 2020 peak).
—Energy and process CO
2
. Three scenarios, each following the same pathway to
a 2020 peak, but subsequently reducing at different rates to maintain total
annual CO
2
e reductions of 3, 5 and 7 per cent.
The following sections detail the deforestation and non-CO
2
greenhouse gas
emission scenarios used to derive the post-peak energy and process CO
2
emission
scenarios and ultimately the total global CO
2
e scenarios for the twenty-first century.
(b)Deforestation emissions
A significant portion of the current global annual anthropogenic CO
2
emissions are attributable to deforestation (in the region of 12–25%). However,
carbon mitigation policy, particularly in OECD nations, tends to focus on those
6
For example, and in particular, the reduced uptake of CO
2
in the Southern Ocean (Raupach et al.
2007) and the potential impact of low level ozone on the uptake of CO
2
in vegetation (Cranmer
et al. 2001).
7
For example, significant uncertainties in the emissions estimates for international shipping
(Corbett & Kohler 2003;Eyring et al. 2005).
5Reframing the climate change challenge
Phil. Trans. R. Soc. A

emissions from energy and industrial processes (hereafter referred to as energy
and process emissions), with less direct regard for emissions arising from defore-
station. While the relatively high levels of uncertainty associated with
deforestation emissions make their inclusion in global mitigation scenarios
problematic, the scale of emissions is such that they must be included. Within
this paper two deforestation scenarios are developed; both assume climate
change to be high on the political agenda and represent relatively optimistic
reductions in the rate of, and hence the total emissions released from,
deforestation.
8
They both have a year 2000 baseline of 5.5 GtCO
2
, but post-
2015 have different deforestation rates and hence different stocks of carbon
remaining in 2100 (i.e. the amount of carbon stored in the remaining forest). The
scenarios are illustrated numerically in table 1 and graphically in figure 1.
The scenarios are dependent not only on the baseline but also on estimates of the
change in forestry carbon stocks between 2000 and 2100. The stock values used in
the scenarios are taken from Moutinho & Schwartzman (2005) and based on their
estimate of total forest carbon stock in 2000 of 1060 GtCO
2
. According to their
assumptions, the carbon stock continues to be eroded at current rates until either
2012 or 2025, following which emissions from deforestation decline to zero by either
2100 or until they equate to 15 per cent of a particular nation’s forest stock
(compared with 2000). They estimate two values for the carbon stocks, released as
CO
2
emissions by 2100 as 319 and 477 GtCO
2
. This implies that within their
scenarios, either 70 or 55 per cent of total carbon stocks remain globally. Given that
this paper and its accompanying AB1 and AB2 scenarios are premised on climate
change being high on the international agenda, Moutinho & Schwartzman’s 55 per
cent of total carbon stock value is considered too pessimistic within the context of
this analysis, and although presented in figure 1, is not included in the analysis
from this point onwards. Moreover, to allow for a more stringent curtailment of
deforestation, the scenario developed for a 70 per cent stock-remaining estimate is
complemented by one with 80 per cent remaining.
Table 1. Deforestation emission scenario summary for two scenarios used to build the subsequent
full CO
2
e scenarios (deforestation low, D
L
; deforestation high, D
H
) and one for illustrative purposes
only (deforestation very high, D
VH
).
name
2000 emissions/
year (carbon
stock) [GtCO
2
] peak date
2100 carbon stock
remaining % (carbon
stock) [GtCO
2
]
emissions 2000–
2100 [GtCO
2
]
D
L
(developed for
this analysis)
5.5 (1060) 2015 80 (847) 213
D
H
(Moutinho &
Schwartzman)
5.5 (1060) 2020 70 (741) 319
D
VH
(Moutinho &
Schwartzman)
5.5 (1060)2036 55 (583)477
8
While the scenarios are at least as optimistic as those underpinning, for example, the 2005 Forest
Resource Assessment (FAO 2005) and the 2006 Stern report, it could be argued they are broadly in
keeping with the high profile deforestation gained during the 2007 United Nations Climate Change
Conference in Bali.
K. Anderson and A. Bows6
Phil. Trans. R. Soc. A

The D
L
and D
H
curves both assume no increase in deforestation rates from
current levels, with D
L
beginning to drop from the peak level of 5.5 GtCO
2
,5
years prior to D
H
. This, combined with the higher level of forestry, and hence
carbon stock remaining in 2100, gives the D
L
curve a faster rate of reduction in
deforestation than is the case for the D
H
curve (typically, 7.4 and 4.8% for D
L
and D
H
, respectively).
9
(c)Non-CO
2
greenhouse gas emissions
To estimate the percentage reductions required from energy and process CO
2
emissions for both AB1 and AB2, it is necessary to consider a range of future
emission scenarios for the non-CO
2
greenhouse gases. Accordingly, three
scenarios are developed assuming current US Environmental Protection Agency
(EPA) estimates and projections of emissions from 2000 up to a range of peaking
years, after which emissions are assumed to decline towards the same long-term
stable level. All the scenarios represent a long-term halving in emission intensity,
with the difference between them arising from the range of cumulative emissions
associated with each of the peaking dates. The scenarios are illustrated
numerically in table 2 and graphically in figure 2.
Anthropogenic non-CO
2
greenhouse gas emissions are dominated by methane
and nitrous oxide and, along with the other non-CO
2
greenhouse gases,
accounted for approximately 9.5 GtCO
2
e in 2000 (EPA 2006; similar figures
are used within the Stern Review), equivalent to 23 per cent of global CO
2
e
2000
0
1
2
3
4
emissions of CO2 per annum (GtCO2)
5
6
7
2020 2040 2060
y
ear
2080 2100
Figure 1. Deforestation emission scenarios showing three CO
2
emissions pathways based on
varying levels of carbon stocks remaining in 2100. Solid curve, 80% stock remaining; dot-dashed
curve, 70% stock remaining; dotted curve, 55% stock remaining.
9
D
L
per cent change value is the mean for the period between 2030 and 2050, and D
H
is the mean
value for 2040–2060.
7Reframing the climate change challenge
Phil. Trans. R. Soc. A

emissions. Understanding how this significant portion of emissions may change in
the future is key to exploring the scope for future emissions reduction from all the
greenhouse gases.
The three non-CO
2
greenhouse gas scenarios presented here are broadly
consistent with a global drive to alleviate climate change. The principal difference
between the scenarios is the date at which emissions are assumed to peak, with the
range chosen to match that for the total CO
2
e emissions, namely an early-action
scenario where emissions peak in 2015, a mid-action peak of 2020 and finally a late-
action peak in 2025. All three scenarios have a growth rate from the year 2000 up
until a few years prior to the peak, equivalent to that projected by the EPA (2006),
10
and broadly in keeping with recent trend data. The scenarios all contain a smooth
transition through the period of peak emissions and on to a pathway leading
towards a post-2050 value of 7.5 GtCO
2
e. This value is again specifically chosen to
2000
0
2
4
6
8
10
12
14
2020 2040 2060
y
ear
2080 2100
emissions of non-CO2 greenhouse gases
per annum (GtCO2e)
Figure 2. Three non-CO
2
greenhouse gas emission scenarios with emission pathways peaking at
different years but all achieving the same residual level by 2050. Short dashed curve, early action;
long dashed curve, mid-action; solid curve, late action.
Table 2. Non-CO
2
greenhouse gas emission scenario summary.
name
2000 emissions
[GtCO
2
] peak year
mean growth
to peak (%)
peak annual
emission
[GtCO
2
e]
total 2000–2100
emissions
[GtCO
2
e]
early action 9.5 2015 1.31 11.4 858
mid-action 9.5 2020 1.51 12.2 883
late action 9.5 2025 1.53 13.3 916
10
EPA values for global warming potential of the basket of six gases are slightly different from
those used in IPCC. The difference, though noted here, does not significantly alter the analysis
or results.
K. Anderson and A. Bows8
Phil. Trans. R. Soc. A

reflect a genuine global commitment to tackle climate change. It is approximately
25 per cent lower than the current level and consistent with a number of other
450 ppmv scenarios.
11
Given that the majority of the non-CO
2
greenhouse gas
emissions are associated with food production, it is not possible, with our current
understanding of theissues, to envisage how emissions could tend to zero while there
remains a significant human population. The 7.5 GtCO
2
e figure used in this paper,
assuming a global population in 2050 of 9 billion (thereafter remaining stable), is
equivalent to approximately halving the emission intensity of current food
production. While a reduction of this magnitude may be considered ambitious in
a sector with little overall emission elasticity, such improvements are necessary if
global CO
2
e concentrations are to be maintained within any reasonable bounds.
The non-CO
2
greenhouse gas scenarios have similar growth rates from 2000 to
their respective peak values, and ultimately all have the same post-2050 emission
level (7.5 GtCO
2
e). The rate of reduction in emissions from the respective peaks
demonstrates the importance of timely action to curtail the current rise in annual
emissions: the early-action scenario is required to reduce at 1.35 per cent per year,
while the mid- and late-action scenario values are at 2 and 3 per cent, respectively.
Similarly, table 2 and figure 2 demonstrate the importance for cumulative values
of non-CO
2
greenhouse gas emissions not rising much higher than todayand that the
post-peak reduction rate achieves the long-term residual emission level as soon as is
possible (7.5 GtCO
2
e by 2050). If the year in which emissions reach the residual level
had been 2100 rather than 2050, the modest differences in cumulative emissions
between the early-, mid- and late-action scenarios would have been substantially
increased. Given that the cumulative value of non-CO
2
greenhouse gas emissions is a
significant proportion of total cumulative CO
2
e emissions, any delay in achieving
the residual value would have significant implications for the reduction rate of
energy and process CO
2
emissions necessary to meet the AB1 and AB2 criteria.
(d)CO
2
e emission scenarios for the twenty-first century
Having developed the deforestation and non-CO
2
greenhouse gas scenarios,
this section presents the complete greenhouse gas emission scenarios, AB1 and
AB2, for the twenty-first century. The emissions released from the year 2000
until the peak dates are discussed here in relation to both AB1 and AB2, before
the post-peak scenarios for each of the scenario sets are presented.
(i) AB1 and AB2: emissions from 2000 to the peak years
By combining the deforestation and non-CO
2
greenhouse gas scenarios with
assumptions about energy and process CO
2
, scenarios for all greenhouse gas
emissions up until the three peaking dates are developed. Energy and process
CO
2
emissions for the years 2000–2005 are taken from the Carbon Dioxide
Information Analysis Centre (CDIAC), with estimates for 2006–2007 based on
BP inventories (BP 2007). From 2007 to the three peaking dates of 2015 (early
action), 2020 (mid-action) and 2025 (late-action) emissions of energy and process
CO
2
grow at 3 per cent per year until 5 years prior to peaking. Beyond this point,
emission growth gradually slows to zero at the peak year before reversing
11
For example, in Stern (2006, p. 233), for both his 450 ppmv CO
2
e and 500–450 ppmv
overshoot curve.
9Reframing the climate change challenge
Phil. Trans. R. Soc. A

thereafter. The 3 per cent emission growth rate chosen for CO
2
is broadly
consistent with recent historical trends. Between 2000 and 2005, CDIAC data
show a mean annual growth in energy and process CO
2
emissions of 3.2 per cent;
this includes the slow growth years following the events of 11 September 2001.
(ii) AB1: emissions from peak years to 2100
From the peak years onwards, AB1 (summarized in table 3) takes the approach that
to remain within the bounds of a 450 ppmv CO
2
e stabilization target, the cumulative
emissions between 2000 and 2100 must not exceed the range presented within the
latest IPCC report in which carbon-cycle feedbacks are included (IPCC 2007b).
(iii) AB1 final scenarios
The emission pathways for the full greenhouse gas AB1 scenarios from 2000 to
2100 are presented in figure 3. The plots comprise the earlier deforestation and
non-CO
2
greenhouse gas scenarios with growing energy and process CO
2
emissions up to the peaking year, and all have total twenty-first century
cumulative values of CO
2
e matching the 450 ppmv figures within AR4.
It is evident from the data underpinning figure 3 that 10 of the 18 proposed
pathways cannot be quantitatively reconciled with the cumulative CO
2
e
emissions budgets for 450 ppmv provided within AR4. Table 4 identifies the
‘impossible’ scenarios (including three with prolonged annual reduction rates
greater than 15%) and illustrates the post-peak level of sustained emission
reduction necessary to remain within budget.
(iv) AB1: implications for energy and process CO
2
The constraints on the greenhouse gas emission pathways of achieving 450 ppmv
CO
2
erendermostoftheAB1 scenarios impossible to achieve. Having established
which scenarios are at least quantitatively possible and subtracting the respective
non-CO
2
greenhouse gas and deforestation emissions, the energy and process emissions
associated with each of the scenarios that remain feasible (figure 4) can be derived.
Figure 4 illustrates that complete decarbonization of the energy and process
system is necessary by between 2027 and 2063, if the total greenhouse gas
emissions are to remain within the IPCC’s 450 ppmv CO
2
e budgets. Moreover, in
combination with table 5, it is evident that the only meaningful opportunity for
Table 3. Summary of the core components of scenario set AB1.
characteristic 2015–2100 2020–2100 2025–2100
deforestation
a
D
H
and D
L
D
H
and D
L
D
H
and D
L
non-CO
2
greenhouse gases
a
early action mid-action late action
approximate peaking value [GtCO
2
e] 54 60 64
cumulative emissions [GtCO
2
e] IPCC
AR4
low: 1376 low: 1376 low: 1376
medium: 1798 medium: 1798 medium: 1798
high: 2202 high: 2202 high: 2202
2100 residual emissions [GtCO
2
e] 7.5 7.5 7.5
a
Deforestation and non-CO
2
greenhouse gas scenarios as in tables 1 and 2.
K. Anderson and A. Bows10
Phil. Trans. R. Soc. A

2000
0
20
40
60
80
2020 2040 2060
year
2080 2100
emissions of greenhouse
gases (GtCO2e)
2000 2020 2040 2060
year
2080 2100
(a)(b)
2000
0
20
40
60
80
2020 2040 2060
y
ear
2080 2100
emissions of greenhouse
gases (GtCO2e)
(c)
Figure 3. Greenhouse gas emission scenarios for AB1 with emissions peaking in (a) 2015, (b) 2020
and (c) 2025. Dark purple curve, low D
L
; black curve, low D
H
; blue curve, medium D
L
; red curve,
medium D
H
; light purple curve, high D
L
; green curve, high D
H
.
Table 4. Scenarios assessed in relation to their practical feasibility. (X denotes a scenario rejected
on the basis of being quantitatively impossible or with prolonged percentage annual reduction rates
greater than 15%. The percentage reductions given illustrate typical sustained annual emission
reductions required to remain within budget.)
peak date
deforestation D
L
deforestation D
H
low medium high low medium high
2015 X 13% 4% X X 4%
2020 X X 8% X X 11%
2025 X X X X X X
11Reframing the climate change challenge
Phil. Trans. R. Soc. A

stabilizing at 450 ppmv CO
2
e occurs if the highest of the IPCC’s cumulative
emissions range is used and if emissions peak by 2015.
(v) AB2: emissions from 2020 (peak year)to 2100
The AB2 scenario set complements the AB1 scenario set by exploring the
implications for CO
2
e budgets of three post-peak annual emission reduction rates
(3, 5 and 7%). Only one peaking year is considered within this analysis with 2020
chosen as arguably the most ‘realistic’ of the three dates in terms of both the
‘practicality’ of being achieved and of the respective scope for remaining within
‘reasonable’ bounds of CO
2
e concentrations. Table 6 summarizes the data
underpinning figure 5.
2000
0
20
40
60
emissions of CO2 (GtCO2)
80
2020 2040 2060
y
ear
2080 2100
Figure 4. Energy and process CO
2
emissions derived by subtracting the non-CO
2
emissions and
deforestation emissions from the total greenhouse gas emissions over the period of 2000–2100, for
the AB1 scenarios. Black solid curve, 2015 peak medium D
L
; black dashed curve, 2015 peak high
D
L
; dot-dashed curve, 2015 peak high D
H
; grey solid curve, 2020 peak high D
L
; grey dashed curve,
2020 peak high D
H
.
Table 5. Twenty-year sustained post-peak per cent reductions in energy and process CO
2
emissions
(from 5 years following the peak year). (X denotes a scenario rejected on the basis of being
quantitatively impossible, with prolonged per cent annual reduction rates greater than 15% or
scenarios where full decarbonization is necessary within 20 years.)
peak date
deforestation D
L
deforestation D
H
low medium high low medium high
2015 X X w6% X X w8%
2020 X X X X X X
2025 X X X X X X
K. Anderson and A. Bows12
Phil. Trans. R. Soc. A

The pathways within figure 5 equate to a range in cumulative CO
2
e emissions for
2000–2100 of 2.4 TtCO
2
e, 2.6 TtCO
2
eand3TtCO
2
e for 7, 5 and 3 per cent reductions,
respectively. According to the cumulative emissions data contained within the Stern
Review (Stern 2006: figure 8.1, p. 222), the first two values approximate to a CO
2
e
concentration of approximately 550 ppmv with the latter being closer to 650 ppmv.
(vi) AB2: implications for energy and process CO
2
Having developed the total CO
2
e pathways for AB2,andgiventhe
deforestation and non-CO
2
greenhouse gas emission scenarios outlined earlier,
the associated energy and process CO
2
scenarios can be derived (figure 6).
Table 7 indicates typical post-peak annual reduction rates in energy and process
CO
2
emissions for the families of 3, 5 and 7 per cent CO
2
e scenarios.
Table 6. Summary of the core components of the AB2 scenarios.
characteristic 2020–2100
deforestation
a
D
H
and D
L
non-CO
2
greenhouse gases
a
mid-action
approximate peaking value [GtCO
2
e] 60
post-2020 CO
2
e reductions (%) 3, 5 and 7
2100 residual emissions [GtCO
2
e] 7.5
a
Deforestation and non-CO
2
greenhouse gas scenarios as in tables 1 and 2.
2000
0
20
40
60
emissions of greenhouse gases (GtCO2e)
80
2020 2040 2060
y
ear
2080 2100
Figure 5. Greenhouse gas emission scenarios peaking in 2020, with sustained percentage emission
reductions of 3, 5 and 7%. The 3 and 5% D
H
scenarios are so similar to the 3 and 5% D
L
that they are
hidden behind those profiles. Black solid curve, 7% reduction D
L
; black dashed curve, 7% reduction
D
H
; thin grey solid curve, 5% reduction D
L
; thin grey dashed curve (hidden), 5% reduction D
H
;thick
grey solid curve, 3% reduction D
L
; thick grey dashed curve (hidden), 3% reduction D
H
.
13Reframing the climate change challenge
Phil. Trans. R. Soc. A

According to these results, the 3, 5 and 7 per cent CO
2
e annual reduction rates
comprising the AB2 scenarios correspond with energy and process decarboniza-
tion rates of 3–4, 6–7 and 9–12 per cent, respectively. While the latter two ranges
correlate broadly with stabilization at 550 ppmv CO
2
e, the former, although
arguably offering less unacceptable rates of reduction, correlates with
stabilization nearer 650 ppmv CO
2
e.
4. Discussion
(a)AB1 scenarios
The AB1 scenarios presented here focus on 450 ppmv CO
2
e and can be broadly
separated into three categories.
(i) Scenarios that quantitatively exceed the IPCC’s 450 ppmv CO
2
e budget
range: this equates to 10 of the 18 scenarios. Scenarios in this category are
quantitatively impossible.
2000
0
20
40
60
CO2 emissions from energy and
industrial processes (GtCO2)
80
2020 2040 2060
y
ear
2080 2100
Figure 6. CO
2
emissions derived by removing the non-CO
2
greenhouse gas emissions and
deforestation emissions from the total greenhouse gas emissions over the period of 2000–2100 for
the AB2 scenarios. Black dashed curve, 7% reduction D
L
; black solid curve, 7% reduction D
H
; thin
grey dashed curve, 5% reduction D
L
; thin grey solid curve, 5% reduction D
H
; thick grey dashed
curve, 3% reduction D
L
; thick grey solid curve, 3% reduction D
H
.
Table 7. Post-peak (2020) per cent reduction in energy and process CO
2
emissions.
annual reduction deforestation D
L
(%) deforestation D
H
(%)
total CO
2
e 35735 7
energy and process CO
2
3694712
K. Anderson and A. Bows14
Phil. Trans. R. Soc. A

(ii) Scenarios with current emission growth continuing until 2015, emissions
peaking by 2020 and thereafter undergoing dramatic annual reductions of
between 8 and 33 per cent. Scenarios in this category are, for the purpose
of this paper, considered politically unacceptable.
(iii) Scenarios that, as early as 2010, break with current trends in emissions
growth, with emissions subsequently peaking by 2015 and declining
rapidly thereafter (approx. 4% per year). Scenarios in this category are
discussed below.
For scenarios within category (iii) to be viable, it is necessary that the IPCC’s
upper value for 450 ppmv cumulative emissions between 2000 and 2100 be
correct. If, on the other hand, the IPCC’s mid- or low value turns out be more
appropriate, category (iii) scenarios will either be politically unacceptable (i.e.
above 8% per annum reduction) or quantitatively impossible.
However, even should the IPCC’s high level (‘optimistic’) value be correct, the
accompanying 4 per cent per year reductions in CO
2
e emissions beginning in under
a decade from today (i.e. by 2018) are unlikely to be politically acceptable without
a sea change in the economic orthodoxy. The scale of this challenge is brought into
sharp focus in relation to energy and process emissions. According to the analysis
conducted in this paper, stabilizing at 450 ppmv requires, at least, global energy
related emissions to peak by 2015, rapidly decline at 6–8 per cent per year between
2020 and 2040, and for full decarbonization sometime soon after 2050.
The characteristics of the resulting 450 ppmv scenario are summarized in
table 8. This assumes that the most optimistic of the IPCC’s range of cumulative
emission values is broadly correct. While this analysis suggests stabilizing at
450 ppmv is theoretically possible, in the absence of an unprecedented step
change in the global economic model and the rapid deployment of successful CO
2
scrubbing technologies, 450 ppmv is no longer a viable stabilization concen-
tration. The implications of this for climate change policy, particularly
adaptation, are profound. The framing of climate change policy is typically
informed by the 28C threshold; however, even stabilizing at 450 ppmv CO
2
e
offers only a 46 per cent chance of not exceeding 28C(Meinshausen 2006). As a
consequence, any further delay in global society beginning down a pathway
towards 450 ppmv leaves 28C as an inappropriate and dangerously misleading
mitigation and adaptation target.
Table 8. Summary of the core components of the 450 ppmv scenario considered theoretically
possible within the constraints of the analysis and assuming the IPCC’s most ‘optimistic’ 450 ppmv
CO
2
e cumulative value.
characteristics quantity
IPCC 450 ppmv upper limit cumulative value for 2000–2100 (GtCO
2
e) 858
peak in CO
2
e emissions 2015
post-peak annual CO
2
e decarbonization rate w4%
total decarbonization date (including forestry and excluding non-CO
2
e
greenhouse gas residual)
w2060–2075
post-peak sustained annual energy and process decarbonization rate w6–8%
total energy and process decarbonization date w2050–2060
15Reframing the climate change challenge
Phil. Trans. R. Soc. A

(b)AB2 scenarios
From the analysis underpinning the AB2 scenarios, it is evident that the rates of
emission reduction informing much of the climate change debate, particularly in
relation to energy, correlate with higher stabilization concentrations than is
generally recognized. The principal reason for this divergence arises, in the first
instance, from the difference between empirical and modelled emissions data for
post-2000. For example, in describing ‘[T]he Scale of the Challenge’ Stern’s
‘stabilization trajectories’ assume a mean annual emissions growth almost 1.5 per
cent lower than was evident from the empirical data between 2000 and 2006. While
the subsequent impact on cumulative emissions for this period is, in itself,
significant, the substantive difference arises from short-term extrapolations of
current trends. Stern’s range of peak emissions for 2015 are some 10 GtCO
2
elower
than would be the case if present trends continued out to 2010, with growth
subsequently reducing to give a peak in emissions by 2015.
12
This substantial
divergence in emissions is exacerbated significantly as the peak date goes beyond
2015. If emissions were to peak by 2020 (as was assumed for the AB2 scenarios), and
again following a slowing in growth during the 5 years prior to the peaking date,
emissions would, by 2020, be between 14 and 16 GtCO
2
e higher than Stern’s 2020
range. This difference alone equates to over a third of current global annual
emissions, with knock-on implications for short- to medium-term cumulative
emissions seriously constraining the viable range of long-term stabilization targets.
While climate change is claimed to be a central issue within many policy
dialogues, rarely are absolute annual carbon mitigation rates greater than 3 per cent
considered viable. In addition, where mitigation polices are more developed, seldom
do they include emissions from international shipping and aviation (Bows &
Anderson 2007). Stern (2006, pp. 231) drew attention to historical precedents of
reductions in carbon emissions, concluding that annual reductions of greater than 1
per cent have ‘been associated only with economic recession or upheaval’. For
example, the collapse of the former Soviet Union’s economy brought about annual
emission reductions of over 5 per cent for a decade. By contrast, France’s 40-fold
increase in nuclear capacity in just 25 years and the UK’s ‘dash for gas’ in the 1990s
both corresponded, respectively, with annual CO
2
and greenhouse gas emission
reductions of only 1 per cent (not including increasing emissions from international
shipping and aviation). Set against this historical experience, the reduction rates
contained within the AB2 scenarios are without a structurally managed precedent.
In all but one of the AB2 scenarios, the challenge faced with regard to total
CO
2
e reductions is increased substantially when considered in relation to
decarbonizing the energy and process systems. Despite the optimistic
deforestation and non-CO
2
greenhouse gas emission scenarios developed for
this paper, the repercussions for energy and process emissions are extremely
severe. Stabilization at 550 ppmv CO
2
e, around which much of Stern’s analysis
12
Comparing values outlined in Stern (2006, p. 233) with those in AB1 and AB2 for 2015. In
addition, Stern envisages a global CO
2
e emissions increase of approximately 5 GtCO
2
e between
2000 and 2015 compared with provisional estimates for China alone of between 4.2 and
5.5 GtCO
2
e, extending up to 12.2 GtCO
2
e (T. Wang & J. Watson of the Sussex Energy Group
(SEG) 2008, personal communication). If the lower SEG estimate for China is correct, Stern’s
analysis implicitly assumes that global emissions (excluding China) remain virtually unchanged
between 2000 and 2015.
K. Anderson and A. Bows16
Phil. Trans. R. Soc. A

revolved, requires global energy and process emissions to peak by 2020 before
beginning an annual decline of between 6 and 12 per cent; rates well in excess of
those accompanying the economic collapse of the Soviet Union. Even for the 3
per cent CO
2
e reduction scenario (i.e. stabilization at 600–650 ppmv CO
2
e), the
current rapid growth in energy and process CO
2
emissions would need to cease
by 2020 and begin reducing at between 3 and 4 per cent annually soon after.
It is important to note that for both AB1 and AB2 scenarios, there is a risk of
a transient overshoot of the ‘desired’ atmospheric concentration of greenhouse
gases as a consequence of the rate of change in the emission pathway. Given
that overshoot scenarios remain characterized by considerable uncertainty and
are the subject of substantive ongoing research (e.g. Schneider & Mastrandrea
2005;Nusbaumer & Matsumoto 2008), they have not been addressed within
either AB1 or AB2.
5. Conclusions
Given the assumptions outlined within this paper and accepting that it considers
the basket of six gases only, incorporating both carbon-cycle feedbacks and the
latest empirical emissions data into the analysis raises serious questions about
the current framing of climate change policy. In the absence of the widespread
deployment and successful application of geoengineering technologies (sometimes
referred to as macro-engineering technologies) that remove and store atmos-
pheric CO
2
, several headline conclusions arise from this analysis.
— If emissions peak in 2015, stabilization at 450 ppmv CO
2
e requires subsequent
annual reductions of 4 per cent in CO
2
e and 6.5 per cent in energy and process
emissions.
— If emissions peak in 2020, stabilization at 550 ppmv CO
2
e requires subsequent
annual reductions of 6 per cent in CO
2
e and 9 per cent in energy and process
emissions.
— If emissions peak in 2020, stabilization at 650 ppmv CO
2
e requires subsequent
annual reductions of 3 per cent in CO
2
e and 3.5 per cent in energy and process
emissions.
These headlines are based on the range of cumulative emissions within IPCC
AR4 (for 450 ppmv) and the Stern report (for 550 and 650 ppmv),
13
with the
accompanying rates of reduction representing the mid-values of the ranges
discussed earlier. While for both the 550 and 650 ppmv pathways peak dates
beyond 2020 would be possible, these would be at the expense of a significant
increase in the already very high post-peak emission reduction rates.
These conclusions have stark repercussions for mitigation and adaptation policies.
By association, they raise serious questions as to whether the current global
economic orthodoxy is sufficiently resilient to absorb the scale of the challenge faced.
13
The 450 ppmv figure is from AR4 (IPCC 2007a), while the 550 and 650 ppmv figures are from
Jones et al. (2006) and include carbon-cycle feedbacks (used in Stern’s analysis). Although the
Jones et al. figures are above the mid-estimates of the impact of feedbacks, there is growing
evidence that some carbon-cycle feedbacks are occurring earlier than was thought would be the
case, e.g. the reduced uptake of CO
2
by the Southern Ocean (Raupach et al. 2007).
17Reframing the climate change challenge
Phil. Trans. R. Soc. A

It is increasingly unlikely that an early and explicit global climate change
agreement or collective ad hoc national mitigation policies will deliver the urgent
and dramatic reversal in emission trends necessary for stabilization at 450 ppmv
CO
2
e. Similarly, the mainstream climate change agenda is far removed from the
rates of mitigation necessary to stabilize at 550 ppmv CO
2
e. Given the
reluctance, at virtually all levels, to openly engage with the unprecedented
scale of both current emissions and their associated growth rates, even an
optimistic interpretation of the current framing of climate change implies that
stabilization much below 650 ppmv CO
2
e is improbable.
The analysis presented within this paper suggests that the rhetoric of 28Cis
subverting a meaningful, open and empirically informed dialogue on climate
change. While it may be argued that 28C provides a reasonable guide to the
appropriate scale of mitigation, it is a dangerously misleading basis for informing
the adaptation agenda. In the absence of an almost immediate step change in
mitigation (away from the current trend of 3% annual emission growth),
adaptation would be much better guided by stabilization at 650 ppmv CO
2
e(i.e.
approx. 48C).
14
However, even this level of stabilization assumes rapid success in
curtailing deforestation, an early reversal of current trends in non-CO
2
greenhouse
gas emissions and urgent decarbonization of the global energy system.
Finally, the quantitative conclusions developed here are based on a global
analysis. If, during the next two decades, transition economies, such as China,
India and Brazil, and newly industrializing nations across Africa and elsewhere
are not to have their economic growth stifled, their emissions of CO
2
e will
inevitably rise. Given any meaningful global emission caps, the implications of
this for the industrialized nations are bleak. Even atmospheric stabilization at
650 ppmv CO
2
e demands the majority of OECD nations begin to make
draconian emission reductions within a decade. Such a situation is unprecedented
for economically prosperous nations. Unless economic growth can be reconciled
with unprecedented rates of decarbonization (in excess of 6% per year
15
), it is
difficult to envisage anything other than a planned economic recession being
compatible with stabilization at or below 650 ppmv CO
2
e.
Ultimately, the latest scientific understanding of climate change allied with
current emission trends and a commitment to ‘limiting average global
temperature increases to below 48C above pre-industrial levels’, demands a
radical reframing
16
of both the climate change agenda, and the economic
characterization of contemporary society.
14
Meinshausen (2006) estimates the mid-range probability of exceeding 48C at approximately 34
per cent for 600 ppmv and 40 per cent for 650 ppmv. Given this analysis has not factored in a range
of other issues with likely net positive impacts, adapting for estimated impacts of at least 48C
appears wise.
15
At 650 ppmv the range of global decarbonization rate is 3–4 per cent per year (table 7, columns 1
and 4). As OECD nations represent approximately 50 per cent of global emissions, and assuming
continued CO
2
emission growth from non-OECD nations for the forthcoming two decades, the OECD
nations will need to compensate with considerably higher rates of emission reductions.
16
This is not assumed desirable or otherwise, but is a conclusion of (i) the quantitative analysis
developed within the paper, (ii) the premise that stabilization in excess of 600–650 ppmv CO
2
e
should be avoided and (iii) Stern’s assertion that annual reductions of greater than 1 per cent have
‘been associated only with economic recession or upheaval’ (Stern 2006, p. 231).
K. Anderson and A. Bows18
Phil. Trans. R. Soc. A

References
Achard, F., Eva, H. D., Mayaux, P., Stibig, H.-J. & Belward, A. 2004 Improved estimates of net
carbon emissions from land cover change in the tropics for the 1990s. Glob. Biogeochem. Cycles
18, GB2008. (doi:10.1029/2003GB002142)
Anderson, K. & Bows, A. 2007 A response to the draft climate change bill’s carbon reduction
targets. Tyndall Centre briefing note 17, Tyndall Centre for Climate Change Research. See
http://www.tyndall.ac.uk/publications/briefing_notes/bn17.pdf.
Andreae, M. O., Jones, C. D. & Cox, P. M. 2005 Strong present-day aerosol cooling implies a hot
future. Nature 435, 1187–1190. (doi:10.1038/nature03671)
Bows, A. & Anderson, K. L. 2007 Policy clash: can projected aviation growth be reconciled with
the UK Government’s 60% carbon-reduction target? Trans. Policy 14, 103–110. (doi:10.1016/
j.tranpol.2006.10.002)
BP 2007 Statistical review of world energy 2007, British Petroleum. See http://www.bp.com/
liveassets/bp_internet/globalbp/globalbp_uk_english/reports_and_publications/statistical_ene
rgy_review_2007/STAGING/local_assets/downloads/spreadsheets/statistical_review_full_rep
ort_workbook_2007.xls.
Canadell, J. G. et al. 2007 From the cover: contributions to accelerating atmospheric CO
2
growth
from economic activity, carbon intensity, and efficiency of natural sinks. Proc. Natl Acad. Sci.
USA 104, 18 866–18 870. (doi:10.1073/pnas.0702737104)
Corbett, J. J. & Kohler, H. W. 2003 Updated emissions from ocean shipping. J. Geophys. Res. 108,
4650. (doi:10.1029/2003JD003751)
Cox, P. M., Huntingford, C. & Jones, C. D. 2006 Conditions for sink-to-source transitions and
runaway feedbacks from the land carbon-cycle. In Avoiding dangerous climate change (eds H. J.
Schellnhuber, W. Cramer, N. Nakicenovic, T. Wigley & G. Yohe), pp. 155–161. Cambridge,
UK: Cambridge University Press.
Cranmer, W. et al. 2001 Global response of terrestrial ecosystem structure and function to CO
2
and
climate change: results from six dynamic global vegetation models. Glob. Change Biol. 7,
357–373. (doi:10.1046/j.1365-2486.2001.00383.x)
DEFRA 2006 Climate change: the UK Programme 2006. Norwich, UK: HMSO, Department of
Food and Rural Affairs.
DEFRA 2007 Draft climate change bill. London, UK: HMSO, Department of Food and Rural Affairs.
DTI 2006 Our energy challenge: securing clean, affordable energy for the long term. London, UK:
HMSO, Department of Trade and Industry.
EPA 2006 Global anthropogenic non-CO
2
greenhouse gas emissions: 1990–2020. Office of
Atmospheric Programs, Climate Change Division, USA Environmental Protection Agency.
European Commission 2007 Limiting global climate change to 2 degrees Celsius: the way ahead for
2020 and beyond. Brussels, Belgium: Commission of the European Communities.
Eyring, V., Ko
¨hler, H. W., van Aardenne, J. & Lauer, A. 2005 Emissions from international
shipping: 1. The last 50 years. J. Geophys. Res. 110, D17305. (doi:10.1029/2004JD005619)
FAO 2000 Global forest resources assessment 2000. FAO forestry paper no. 140. Food and
Agriculture Organisation of the United Nations, Rome.
FAO 2005 Global forest resources assessment 2005. Global synthesis FAO forestry paper no. 124.
Food and Agriculture Organisation of the United Nations, Rome.
Friedlingstein, P. et al. 2006 Climate-carbon-cycle feedback analysis, results from the C4MIP
model intercomparison. J. Clim. 19, 3337–3353. (doi:10.1175/JCL13800.1)
IPCC 1996 Climate change 1995: the science of climate change. Contribution of working group 1 to
the second assessment report of the Intergovernmental Panel on Climate Change.
IPCC 2007aClimate change 2007: synthesis report. Fourth assessment report of the
Intergovernmental Panel on Climate Change.
IPCC 2007bClimate change 2007: the physical science basis. Report of working group 1 to the
fourth assessment report of the Intergovernmental Panel on Climate Change.
19Reframing the climate change challenge
Phil. Trans. R. Soc. A

Jones, C. D., Cox, P. M. & Huntingford, C. 2006 Impact of climate-carbon-cycle feedbacks on
emissions scenarios to achieve stabilisation. In Avoiding dangerous climate change (eds H. J.
Schellnhuber, W. Cramer, N. Nakicenovic, T. Wigley & G. Yohe), pp. 323–331. Cambridge,
UK: Cambridge University Press.
Le Que
´re
´,C.et al. 2007 Saturation of the Southern Ocean CO
2
sink due to recent climate change.
Science 316, 1735–1738. (doi:10.1126/science.1136188)
Mannstein, H. & Schumann, U. 2005 Aircraft induced contrail cirrus over Europe. Meteorologische
Zeitschrift 14, 549–554. (doi:10.1127/0941-2948/2005/0058)
Meinshausen, M. 2006 What does a 2C target mean for greenhouse gas concentrations? A brief
analysis based on multi-gas emission pathways and several climate sensitivity uncertainty
estimates. In Avoiding dangerous climate change (eds H. J. Schellnhuber, W. Cramer, N.
Nakicenovic, T. Wigley & G. Yohe), pp. 253–279. Cambridge, UK: Cambridge University
Press.
Moutinho, P. & Schwartzman, S. (eds) 2005 Tropical deforestation and climate change. Bele
´m,
Brazil: Amazon Institute for Environmental Research.
Nusbaumer, J. & Matsumoto, K. 2008 Climate and carbon-cycle changes under the overshoot
scenario. Glob. Planet. Change 62, 164–172. (doi:10.1016/j.gloplacha.2008.01.002)
Raupach, M. R., Marland, G., Ciais, P., Le Quere, C., Canadell, J. G., Klepper, G. & Field, C. B.
2007 Global and regional drivers of accelerating CO
2
emissions. Proc. Natl Acad. Sci. USA 104,
10 288–10 293. (doi:10.1073/pnas.0700609104)
RCEP 2000 Energy—the changing climate, 22nd report, CM 4749. The Stationery Office, London.
Schneider, S. H. & Mastrandrea, M. D. 2005 Inaugural article: probabilistic assessment of
“dangerous” climate change and emissions pathways. Proc. Natl Acad. Sci. USA 102,
15 728–15 735. (doi:10.1073/pnas.0506356102)
Stern, N. 2006 Stern review on the economics of climate change. Cambridge, UK: Her Majesty’s
Treasury, Cambridge University Press.
Stordal, F., Myhre, G., Arlander, W., Svendby, T., Stordal, E. J. G., Rossow, W. B. & Lee, D. S.
2004 Is there a trend in cirrus cloud cover due to aircraft traffic? Atmos. Chem. Phys. Discuss. 4,
6473–6501.
Tol, R. S. J. 2007 Europe’s long-term climate target: a critical evaluation. Energy Policy 35,
424–432. (doi:10.1016/j.enpol.2005.12.003)
K. Anderson and A. Bows20
Phil. Trans. R. Soc. A