Regional Carbon Budgets - The Belfer Center for Science and

Transcription

Regional Carbon Budgets - The Belfer Center for Science and
The Harvard Project on Climate Agreements
October 2015
Discussion Paper 15-78
Regional Carbon Budgets:
Do They Matter for Climate
Policy?
Massimo Tavoni
Fondazione Eni Enrico Mattei (FEEM)
Politecnico di Milano
Detlef P. van Vuuren
PBL Netherlands Environmental Assessment Agency
Utrecht University
Email: [email protected]
Website: www.belfercenter.org/climate
Regional Carbon Budgets:
Do They Matter for Climate Policy?
Massimo Tavoni
Fondazione Eni Enrico Mattei (FEEM)
Politecnico di Milano
Detlef P. van Vuuren
PBL Netherlands Environmental Assessment Agency
Utrecht University
Prepared for
The Harvard Project on Climate Agreements
THE HARVARD PROJECT ON CLIMATE AGREEMENTS
The goal of the Harvard Project on Climate Agreements is to help identify and advance scientifically
sound, economically rational, and politically pragmatic public policy options for addressing global
climate change. Drawing upon leading thinkers in Argentina, Australia, China, Europe, India, Japan,
and the United States, the Project conducts research on policy architecture, key design elements, and
institutional dimensions of domestic climate policy and a post-2015 international climate policy
regime. The Project is directed by Robert N. Stavins, Albert Pratt Professor of Business and
Government, Harvard Kennedy School. For more information, see the Project’s website:
http://belfercenter.ksg.harvard.edu/climate.
Acknowledgements
The Harvard Project on Climate Agreements is grateful for support from the Belfer Center for
Science and International Affairs and the Hui Research Fund for Generating Powerful Ideas at the
Ash Center for Democratic Governance and Innovation—both located at the Harvard Kennedy
School; the Harvard University Center for the Environment; Christopher P. Kaneb (Harvard AB
1990); and the International Emissions Trading Association (IETA).
Previous sponsors of the Harvard Project on Climate Agreements include: ClimateWorks
Foundation, the Doris Duke Charitable Foundation, and the James M. and Cathleen D. Stone
Foundation.
The closely affiliated, University-wide Harvard Environmental Economics Program receives
additional support from the Enel Endowment for Environmental Economics at Harvard University,
the Enel Foundation, the Alfred P. Sloan Foundation, the Mossavar-Rahmani Center for Business
and Government at the Harvard Kennedy School, Bank of America, BP, Castleton Commodities
International LLC, Chevron Services Company, Duke Energy Corporation, and Shell.
Citation Information
Tavoni, Massimo, and Detlef P. van Vuuren. “Regional Carbon Budgets: Do They Matter for
Climate Policy?” Discussion Paper 2015-78. Cambridge, Mass.: Harvard Project on Climate
Agreements, October 2015.
The views expressed in the Harvard Project on Climate Agreements Discussion Paper Series are
those of the author(s) and do not necessarily reflect those of the Harvard Kennedy School or of
Harvard University. Discussion Papers have not undergone formal review and approval. Such
papers are included in this series to elicit feedback and to encourage debate on important public
policy challenges. Copyright belongs to the author(s). Papers may be downloaded for personal use
only.
Regional carbon budgets: do they matter for climate policy?
Massimo Tavoni (FEEM and Politecnico di Milano), Detlef P. van Vuuren (PBL and Utrecht University)
Abstract
Carbon budgets have emerged as a robust metric of warming, but their application to climate policy has
been limited to global assessments. This article explores the potential of regional carbon budgets to inform
climate policy. Using the large database of model scenarios from IPCC AR5 WGIII, we show that regional
cumulative emissions are important metrics of the long term contribution to climate change and of the
effort required to mitigate it. Their value appears to be more limited for informing short term courses of
actions, and for predicting the economic consequences of emission reduction policies.
1. Motivation and contribution
Carbon budgets are a powerful way to translate the complexities of climate science into an easily
understandable, yet remarkably robust, linear relation between cumulative emissions and temperature
increase. Since their inception, they have received increasing attention and have been validated in many
different contexts (Zickfeld et al. 2009; Allen et al. 2009; Meinshausen et al. 2009; Matthews et al. 2009),
most notably in the latest assessment report of the IPCC. The relationship derived in the IPCC report on the
basis of climate model calculations of the Representative Concentration Pathways (van Vuuren et al. 2011)
indicate that in order to likely remain below the 2oC target – currently discussed in international
negotiations - the carbon budget should be around 1000 GtCO2 or lower (Friedlingstein et al. 2014).
The proportionality between the sum of CO2 emissions and temperature allows disaggregating across
geography and time: the future global temperature increase can be retrieved by summing up the regional
temperature contributions obtained by direct multiplication of the regional carbon budgets. This allows for
a clear attribution of future warming to individual regions and countries, in the same way historical
emissions provide national responsibility to date (Matthews et al. 2014). Cumulative emissions at the
regional level could also be used to inform mitigation policy, since they share some of the advantages of
the global carbon budget, namely that they allow us to overcome the myopic view of emission quotas in a
given year only, and to establish scientific clarity in the long term (Frame et al. 2014). In spite of these
advantages, so far indicators of cumulative emissions have not been put forward in the policy context.
Most of the proposed mitigation targets, such as the ones of the Intended Nationally Determined
Contributions (INDCs), are expressed in terms of emission reductions in specific years (e.g. 2030), or based
on even less scientifically valid metrics such as year of peaking of emissions. This makes it difficult to judge
the long term temperature impacts of national policies, since a given temperature objective can be attained
through different intermediate mitigation strategies.
However, some countries -such as the EU with its 2050 roadmap- have provided medium as well as longer
term targets (e.g. 2020, 2030 and 2050), which can be translated into cumulative emission indicators
through backcasting and modeling work1. Integrated assessment models have been designed to relate
short-term mitigation actions to long-term climate objectives, as summarized in their contribution to the
IPCC 5th assessment report. The output of these models can thus be used to bridge the gap between short
1
This is indeed the approach taken by the EU commission, see
http://www.roadmap2050.eu/attachments/files/Volume1_fullreport_PressPack.pdf
1
term targets –such as the ones currently proposed as INDCs- and stock indicators such as cumulative
emissions, which are needed to evaluate climate impacts as well as comparability of effort (Aldy & Pizer
2014).
This paper contributes to an assessment of the usefulness of regional cumulative emission indicators to
inform climate policy. So far, most of the literature connecting regional contributions to global targets has
focused on specific annual targets or emissions pathways either using burden sharing rules or assuming a
global cost-effective response. Examples of such approaches include (den Elzen & Höhne 2008; Ciscar et al.
s.d.; Jacoby et al. 2009; Kober et al. 2013; Miketa & Schrattenholzer 2006). An overview of these
approaches is provided by Hof et al. (Hof et al. 2009). Some recent contributions have discussed ways to
sharing cumulative emissions at the regional level (Tavoni et al. 2014; Raupach et al. 2014; Gignac &
Matthews 2015; Anderson et al. 2008). The aim of this paper is to go beyond the existing work and explore
the correlation between cumulative emissions and other policy relevant indicators of mitigation effort.
Earlier, Raupach et al looked at the relationship between emission quotas and the emission reduction rate,
while Tavoni et al looked at regional cumulative emissions. Our work, however, considers more policy
categories and a broader set of indicators, and uses actual outcomes of optimal mitigation strategies as
devised by integrated assessment models, rather than relying on exogenously determined trajectories.
2. Concepts and methods
While the concept of a fixed amount of cumulative emissions, or a carbon budget, is clearly defined at the
global level, when discussing regional budgets, it is important to distinguish between emission allocations
and actual emissions.
Emission allocations, also known as quotas or permits, refer to endowments of emission credits which
regions receive based on some burden sharing principle, and which can be freely exchanged on the market.
Actual emissions refer to what is really emitted by the country. These two concepts coincide in the absence
of a carbon trading market (such as for example in the case of a global carbon tax), but differ whenever
trading of emission credits is allowed across regions. For the purpose of our analysis, it is important to
realize that in policy settings aimed at achieving global targets at the minimum global costs, climate policies
are implemented either through a uniform carbon tax, or via a cap and trade system with a single price on
carbon and trade of CO2 permits across regions. In such a setting, regional emissions are determined by the
regional mitigation potentials (in such a way as to equalize marginal abatement costs): allowances above or
below these optimal values would then be traded (e.g either sold or bought respectively), but would not
have any impact on actual emissions. In this framework, which is the one commonly used by Integrated
Assessment Models (IAMs), emission allowances determine who pays for mitigation, and thus equity and
efficiency can be disentangled (Luderer et al. 2012). This has important economic consequences, which we
take into account, but (at least in the ideal model world) does not matter for actual emissions: once a
carbon price is in place, regional budgets are univocally determined, irrespective of emission allowances. In
this article, we will focus on these actual emissions, as they relate directly to the mitigation effort of the
energy and land use systems. However, when quantifying the economic implications of climate policy, we
deduct revenues or expenditures coming from sales or purchases of emission allowances, in order to
determine the domestic policy costs
In order to explore the potential and limitations of regional budgets to inform climate policy, we use the
large ensemble of scenarios generated by several integrated assessment models (IAMs) for the WGIII of the
2
5th assessment of the IPCC. The database, which is publicly available2, contains several hundred scenarios,
spanning a wide range of climate categories, policy and technology implementations, many of which have
been generated in the context of multi model comparison projects. A great deal of information, including
CO2 and non-CO2 emission pathways, energy system and economic indicators, is contained in the data base.
Most of the climate policy scenarios have focused on long term forcing targets, though some model
ensembles have directly implemented global budgets as policies (Kriegler et al. 2014).
Our calculations consider only the IAMs which can generate long term, i.e. meaning up to 2100, CO2
emissions profiles. In order to span different futures, we have included both ‘first best’ –idealized cases of
full cooperation and full technology availability- and ‘second best’ scenarios – cases in which policy is
fragmented or technology is limited, though we perform sensitivity to the scenario category. In first best
scenarios, which represent the majority of the scenarios reviewed by the IPCC, policies are implemented in
a globally cost effective way by equalizing marginal abatement costs across regions, ensuring a levelized
playfield among regions. In fragmented scenarios, carbon prices can vary across regions, though they tend
to converge over time. We compute results for all 5 IPCC regions, namely OECD, ASIA, Latin and Central
America (LAM), Middle East and Africa (MAF) and REF (Reforming Economies, also known as Economies in
Transition)3. We report the results of the five regions separately: given the regional interdependencies
through energy markets, trade, etc., these should be interpreted consistently with the policy architecture
implemented by IAMs, which as already noted for the most part is that of uniform climate policies.Though
the terms “cumulative emissions” and “carbon budgets” are quite interchangeable, we will use “carbon
budgets” only when discussing climate policy scenarios.
3. Applications of regional cumulative emissions and carbon budgets
We first focus on the results for projected regional cumulative emissions for the no climate policy or
business as usual (BAU) scenarios. We subsequently report on regional budget for policies aimed at
achieving climate stabilization at some predetermined levels.
3.1. No climate policy scenarios
Baseline scenarios (counterfactual scenarios assuming no climate policies) are important because they
quantify the regional contribution to climate change, and because they form the benchmark against which
climate policies are compared. They depend on assumptions about economic growth, fossil fuel availability,
energy intensity change, and low carbon technology potential, which vary within and across IAMs.
Figure 1 reports the baseline CO2 cumulative emissions from 2010 to 2100, for the five representative
regions of interest. The figure shows that in the absence of climate policies, fossil fuels would be sufficiently
abundant and cost competitive to generate large amounts of cumulative emissions.
The median (across IAMs) cumulative emissions of Asia exceed 2500 GtCO2. This budget alone would be
more than double the global allowable budget compatible with 2oC, represented by the colored areas in the
chart. In fact, Asia alone would add more CO2 to the atmosphere in the remaining part of this century than
all the CO2 added since pre-industrial times globally. This is of course a result of the sheer size –population
and economic wise- of the continent. However, large contributions are also expected from other regions.
2
At https://secure.iiasa.ac.at/web-apps/ene/AR5DB/dsd?Action=htmlpage&page=about
See https://secure.iiasa.ac.at/web-apps/ene/AR5DB/dsd?Action=htmlpage&page=about for an exact definition of
the regional aggregates.
3
3
Most notably, the OECD countries, which have already contributed to the largest part of the historical CO2
emissions, are expected to emit in excess of 1000 GtCO2, if no specific policy to reduce emissions were to
be implemented. The individual cumulative emissions for the LAM, REF and potentially the MAF range
could be within the global budget. The sum, however, clearly will not be within this budget. The variations
across models, and also within the same model but for different scenarios, are reflected in the large ranges
reported in Figure 14.
Figure 1. Boxplot of regional cumulative CO2 emissions for Business as usual (BAU) scenarios. On each
box, the central mark is the median, the edges of the box are the 25th and 75th percentiles, the whiskers
extend to 1.5 the interquartile range. The green and red shades indicate the temperature carbon budgets
from IPCC WGIII consistent with 66% and 50% chances of keeping temperature below 2C respectively.
The numbers represent the median carbon budgets (in GtCO2) and temperature increase over 2010
corresponding to an average TCRE of 0.48C/1000GtCO2.
Given the linearity between cumulative emissions, one can also translate the regional cumulative emissions
into regional contributions to peak warming, and subsequently into total warming. The temperature
increase associated with the median budgets, and using a central Transient Climate Response to cumulative
Emissions (TCRE) estimate of 0.48°C/1000GtCO25, is also reported in Figure 1. With this parametrization,
OECD90 and Asia together would add about 1.6°C, in addition to the current warming of 0.7°C. Another
4
It should be remarked, however, that the BAU scenarios in the IPCC DB are not meant to span the full ranges of
possible futures, and thus represent only a subset of potential outcomes of no policy cases. The new shared socio
economic pathways, which have been recently released, have provided additional alternatives, further enlarging the
space of BAU outcomes.
5
This corresponds to 1.75°C/1000GtC, the value adopted in (National Academy Press 2011), which spans the best
estimates of 1.5 of (Matthews et al. 2009) and 2 of (Allen et al. 2009; Zickfeld et al. 2009).
4
0.8°C would come from the remaining regions, LAM, MAF and REF. Additional warming would be generated
by any emission generated after 2100, as well as by non-CO2 gases.
3.2. Climate stabilization scenarios
When a climate stabilization policy is in place, emissions are reduced across the participating regions. In socalled “cost-optimal” scenarios, models assume that emissions are reduced in all regions so that marginal
abatement costs are equalized across these regions. This can be done by emission trading instruments or
by implementing a global carbon tax. Regional carbon budgets consistent with a global climate target can
then derived by summing over time regional actual emissions. This is of course a rough simplification of the
real policy debate, that often discusses targets as deviations from business as usual, and considers a much
larger set of policy instruments. Still, regional carbon budgets could be used as reference to determine
cost-optimal scenarios. These carbon budgets are obviously different from budgets that could be derived
on the basis of fairness principles (e.g. equal per capita emissions) as discussed in Raupach et al. (2015).
Figure 2 reports the cost-optimal regional CO2 budgets for two classes of climate stabilization targets of
different stringency, 430-530 ppm-eq and 530-650 ppm eq respectively6. For each region, the three bars
show the budgets till 2030, 2050, and 2100. Focusing first on the 2100 budgets (rightmost bars for each
region), the chart shows that under climate objectives consistent with the 2oC target, OECD and ASIA would
have budgets around 300 and 600 GtCO2. Compared to the much larger BAU budgets of Figure 1, these
imply that very significant mitigation efforts would be needed in all regions in order to attain climate
stabilization. Figure 2 also highlights important regional differences: LAM, MAF and REF regions have
significantly lower budgets than the OECD and ASIA regions, even compared to the baseline. The larger
relative reduction is a result of different mitigation opportunities across regions. Integrated assessment
models foresee large biological mitigation potential in tropical regions such as LAM. REF and MAF, through
forest management and bioenergy practices (Clarke L., K. Jiang, K. Akimoto, M. Babiker, G. Blanford, K.
Fisher-Vanden, J.-C. Hourcade, V. Krey, E. Kriegler, A. Löschel, et al. s.d.).
6
These correspond to climate categories I and II, and III and IV respectively, as defined in IPCC AR5 WGIII.
5
430-530 ppme
1400
1400
1200
1200
2100
2050
1000
800
-200
-200
REF
REF
0
OECD90
0
MAF
200
LAM
200
ASIA
400
OECD90
400
MAF
600
LAM
GtCO2
800
2030
600
ASIA
1000
GtCO2
530-650 ppme
Figure 2. Boxplot of regional CO2 budgets for two classes of climate stabilization targets (430-530 and
530-650 ppm). For each region, the three boxplots show the budgets from 2010 to 2030, 2050 and 2100
respectively. On each box, the central mark is the median, the edges of the box are the 25th and 75th
percentiles, the whiskers extend to 1.5 times the interquartile range, and outliers are plotted
individually.
A less stringent climate target of 530-650 ppm CO2-e (keeping temperature increase likely below 3oC
warming) leads to higher regional budgets. The global increase of 100 ppm translates into roughly 200
additional GtCO2 of budget for each of the five analyzed regions.
Models also allow us to compute carbon budgets for different moments in time. Figure 2 provides the
additional information of the distribution of the budgets over three policy relevant periods: 2010-2030,
2010-2050, and the already discussed 2010-2100. Several insights emerge. Most of the budget appears to
be concentrated in the early decades of the century, e.g. up to 2030. The reason for this temporal
allocation of the century scale budget is that IAMs take into account the inertia in transforming the energy
system from a fossil fuel based one toward a low carbon one. Also, the reviewed scenarios include cases of
slow progress and delayed action on the front of emission reductions. For many regions, the 2050 budget
appears to be very close to the 2100 one, especially for the most stringent climate category. In some
regions (e.g. LAM), the 2050 budget can be even higher than the 2100 one. The reason is that cumulative
CO2 emissions in the second part of the century in most stringent scenarios are very low or even net
negative. In the LAM region, reforestation policies could lead to even lower emissions (or more negative)
than in other regions. When moving to the less stringent target of 530-650 ppm CO2-e budgets are spread
more evenly over time, thanks to the larger overall headroom. Once again LAM shows a particularly striking
pattern, with no emission growth post 2050. The ranges across models also increase over time, with limited
variance in 2030, and larger ones in 2050 and especially 2100, reflecting diverging projections across IAMs
as time goes by.
6
The issue of negative emissions deserves further scrutiny. IAMs feature mitigation technologies which can
absorb CO2 from the atmosphere, and resort to these when confronted with stringent targets, or even with
lenient climate targets but with delayed mitigation action in the next few decades or with limited
availability of conventional technology. Carbon dioxide removal is thus a key mitigation option under
certain conditions, and most IAMs implement it mostly in terms of biological removal coupled with carbon
capture and storage (ie. BECCS) (Azar et al. 2010; Vuuren et al. 2013; Kriegler et al. 2013; Edmonds et al.
2013; Tavoni & Socolow 2013). The feasibility of large scale negative emissions programmes is hard to
assess at the moment, and will require significant technological progress to become viable (Fuss et al. 2014;
Smith & Torn 2013).
Figure 3 reports the ‘negative carbon budgets’ at the regional level7. They are the cumulative sum (in
absolute values) over the entire century of CO2 emissions during periods of net negative emissions. These
negative carbon budgets explain the patterns observed in Figure 2, with limited or even negative growth of
emissions post 2050. The chart points to significant quantities of net negative emissions, especially in some
regions and for the most stringent climate objective. The median negative emission budget in LAM for the
430-530 target is in the order of 75 GtCO2. Globally, the regional budgets add to several hundred GtCO2 of
net negative emissions. As IAMs generally assume that some residual emissions will remain positive
throughout the century in specific sectors or for certain activities, the negative budgets are smaller than the
total use of carbon dioxide removal (CDR). Some of the global IAMs show cumulative carbon dioxide
removal of up to 1000 GtCO2 (Tavoni & Socolow 2013).
530-650 ppme
350
300
300
250
250
0
0
REF
OECD90
50
MAF
50
LAM
100
ASIA
100
7
REF
150
MAF
150
200
LAM
200
ASIA
GtCO2
350
OECD90
GtCO2
430-530 ppme
This figure focuses on those scenarios which do produce negative emissions. Thus, it is a subsample of the scenarios
shown in Figure 2.
7
Figure 3. ‘Negative CO2’ budgets. Total cumulative emissions during the period of net negative emissions.
On each box, the central mark is the median, the edges of the box are the 25th and 75th percentiles, the
whiskers extend to 1.5 times the interquartile range, and outliers are plotted individually.
The charts shows a great deal of uncertainty over the amount of negative CO2 budgets. This is the outcome
of two processes. First, negative emissions are very sensitive to the policy setting. They play a fundamental
role in scenarios with delayed global participation, fragmented regional action, and limited availability of
conventional mitigation options such as renewables and nuclear power. At the same time, they require
specific technologies. In models, IAMs represent negative emission technologies in such form as CCS: given
the uncertainty around CCS several scenarios in the IPCC DB explored cases without it. Secondly, different
IAMs make importantly different assumptions about the technical and economic potential of CDR, and their
repercussions on land use.
The overall picture suggests that large negative CO2 budgets are an important –albeit uncertaincomponent of the mitigation strategy foreseen by IAMs. For policy making, this has deep repercussions:
carbon budgets of 1000 GtCO2 without negative emissions or combined with 500 of negative CO2 budget
post 2050 are identical in terms of the cumulative emissions across the century, but entail completely
different temporal allocation of emission reductions, transformation of the energy sector, land use change,
etc.
Finally, we take a look at the distinction between carbon budgets (based on actual emissions) and emission
allowances (see Section 2 for the definition). In those policy settings which allow trading of CO2 permits,
significant quantities of CO2 might be exchanged between countries. The magnitude and direction of trade
is determined by the allocation of emission allowances as well as the regional carbon budgets, the latter
being the outcome of the cost-effective repartition of mitigation. The allocations can be set at any level,
e.g. incorporating different assumptions about equity. A common allocation scheme is based on the equity
principle of equalizing per capita emissions across countries, but many others exist (den Elzen & Höhne
2008; Elzen et al. 2012). Although the mitigation strategy –in terms of energy and land use sector
transformation- is solely determined by the regional carbon budgets, the economic consequences are not.
It is indeed the scope of carbon trading to distinguish who will mitigate from who will pay.
The traded CO2 budgets are shown in Figure S1, for the first half and full century respectively. The chart
shows large ranges across models and scenarios, due to the different choices of allocation schemes across
scenarios. LAM and MAF are general sellers of permits, given their relatively low per capita emissions.
Temporal and policy dynamics are present, with some regions –such as the OECD- turning from buyers to
sellers of cumulative emissions over time, and with less stringent climate targets showing higher volumes of
trade. Economic transfers are determined by the carbon price at which permits are exchanged, which will
depend positively on the stringency of the climate target. Previous research has indicated that the trade
flows would be sufficient to finance large portions of clean energy investments in developing regions, but
that the institutional requirements for managing such large markets would be very significant (Wara 2007;
Tavoni et al. 2014).
4. Are regional carbon budgets a good indicator of policy?
8
In the previous section we have shown that cumulative emissions and carbon budgets are useful indicators
for determining both the regional contribution to climate change in no policy (BAU) scenarios, and the
regional stringency of mitigation compatible with given global climate stabilization objectives respectively.
In this section, we take a closer look at the correlation between regional carbon budget and policy effort.
4.1. Correlation with mitigation effort
Cumulative mitigation (%)
Cumulative mitigation (%)
As indicator of policy effort, we first look at the relative reduction in emissions with respect to BAU. This is
one of the most important drivers of energy and carbon intensities, carbon prices and mitigation costs.
Figure 4 shows the relation between regional carbon budgets and cumulative mitigation, both expressed
until 2100. The chart shows a strong correlation between these two indicators. As expected the lower the
budget, the higher the cumulative emission reduction with respect to a scenario without climate policy. In
accordance with what shown in the previous sections, the LAM and MAF regions can accommodate net
negative budgets, which require cumulative mitigation efforts exceeding 100%. Across our five regions of
interest, cumulative mitigation appears to be negatively related to budgets, with correlation indexes
exceeding -0.75.
120
100
120
Corr = -0.78***
100
120
Corr = -0.82***
100
80
80
80
60
60
60
40
40
40
0
500
1000
0
OECD90
120
100
500
1000
-100
ASIA
120
Corr = -0.75***
100
100
80
60
60
60
40
40
40
MAF
400
100 200
120
Corr = -0.81***
80
200
0
LAM
80
0
Corr = -0.91***
-100 0
100 200 300
REF
Corr = -0.86***
1000
2000
World
Figure 4. Relation between regional CO2 budgets (2010-2100) and cumulative mitigation (2010-2100,
relative to BAU), for two groups of climate categories (430-530 in green and 530-650 ppm eq in red). Dots
represent model scenarios. Budgets below 0 and cumulative mitigation above 100% are possible due to
negative emissions. The blue lines show the outcome of quantile regressions, at 5 (dotted), 50 (solid) and
95 (dotted) percentiles. Pearson’s correlation coefficient is reported along with the confidence interval
(***=0.01, **=0.05,*=0.1).
The strong long term relation between budgets and cumulative mitigation is not an obvious one, given that
mitigation is measured against emissions in a counterfactual scenario (BAU), which can vary significantly
across countries (Blanford et al. 2012). Focusing on the shorter term, such as the first half of the century,
9
might be more policy relevant, but -as shown in Figure S2- would weaken the relation between regional
carbon budgets and cumulative emission reduction.
Albeit relevant, cumulative mitigation with respect to BAU is not a common metric for designing short to
medium term climate policies. This is because it does not provide clearly defined targets in specific periods
of time, and also because of the arbitrariness of counterfactual BAU scenarios, which are subject to a great
deal of uncertainty. A more common, though less precise, metric of effort is the mitigation in a given year,
with respect to some given level, e.g. today’s emissions. The INDCs which countries are submitting to the
UNFCCC are often expressed in terms of 2030 emission reductions compared to a reference year. In the
next chart, we explore the relation between emission reductions in 2030, and the regional carbon budget
from 2010 to 2050 –which is known to be already a good indicator of future warming (Meinshausen et al.
2009).
Figure 5 indicates a relatively strong relation between carbon budget and this indicator of mitigation effort,
especially at the global level. Regionally, it appears to be somewhat weaker with respect to what shown in
Figure 4. This is particularly true for the regions which have the smallest budgets, and therefore achieve the
more stringent mitigation, LAM and MAF. The relation is further weakened when focusing on delayed or
fragmented policy scenarios (see Figure S4).
Mitigation (%)
100
100
Corr = -0.82***
50
50
50
0
0
0
-50
-50
-50
200
400
600
400
OECD90
100
Mitigation (%)
100
Corr = -0.78***
Corr = -0.59***
600
0
ASIA
100
100
Corr = -0.67***
50
50
0
0
0
-50
-50
-50
100
MAF
200
0
100
REF
100
200
LAM
50
0
Corr = -0.29***
200
Corr = -0.87***
500 1000 1500 2000
World
Figure 5 Relation between regional CO2 budgets (2010-2050) and mitigation (in 2030, relative to 2010),
for two groups of climate categories (430-530 in green and 530-650 ppm eq in red). Dots represent model
scenarios. The blue lines shows the outcome of quantile regressions, at 5 (dotted), 50 (solid) and 95
(dotted) percentiles. Pearson’s correlation coefficient is reported along with the confidence interval
(***=0.01, **=0.05,*=0.1).
10
As a final exploration, we evaluate the correlation between regional carbon budgets and the timing of
mitigation effort. Both the year reaching negative values or attain a maximum emission level are used as
focal points for climate policy. In the first case, this indicates by when during this century, if ever, the entire
energy and land use system is expected to reach overall carbon neutrality. The second provides an
indication of the time by when emissions will have to begin to decline, which is an important turning point
for those economies where emissions are growing particularly rapidly.
First Year of Neg CO2
Figure 6 shows a relatively clear and strong correlation between regional carbon budgets and the year by
when CO2 emissions are projected to become negative. Scenarios with low or negative regional carbon
budgets are consistent with emissions turning negative shortly after mid century.
2100
2100
2100
2080
2080
2080
2060
2060
2060
2040
2040
Corr = 0.72***
0
500
0
500
OECD90
First Year of Neg CO2
2040
Corr = 0.55***
1000
-100
ASIA
2100
2100
2080
2080
2080
2060
2060
2060
2040
Corr = 0.71***
0
200
MAF
400
2040
Corr = 0.64***
0
100
REF
0
100 200
LAM
2100
2040
Corr = 0.77***
200
Corr = 0.44***
500 1000 1500 2000
World
Figure 6. Relation between regional CO2 budgets and first year of net negative CO2 emissions, for two
groups of climate categories (430-530 in green and 530-650 ppm eq in red). Dots represent model
scenarios. The blue lines show the outcome of quantile regressions, at 5 (dotted), 50 (solid) and 95
(dotted) percentiles. Pearson’s correlation coefficient is reported along with the confidence interval
(***=0.01, **=0.05,*=0.1). The green markers are bigger for improved clarity.
The year of emission peak also has some correlation with budgets, , but only for some countries such as
ASIA (Figure 7). Climate policies consistent with 2oC (430-530 ppm-eq) suggest a peaking of CO2 emissions
in ASIA which would not exceed 2030, and median century scale budgets of about 500 GtCO2. This figure
suggests that the commitment recently made by China to have emissions peak by 2030 is not consistent
with the 2oC target and most importantly would need to be matched by other large Asian economies, going
beyond actions currently discussed in the context of the INDCs.
11
Peak Year
2060
2060
2060
2050 Corr = 0.46***
2050 Corr = 0.43***
2050 Corr = 0.032***
2040
2040
2040
2030
2030
2030
2020
2020
2020
2010
0
500
2010
0
500
Peak Year
OECD90
1000
2010
-100
ASIA
0
100 200
LAM
2060
2060
2060
2050 Corr = 0.31***
2050 Corr = -0.045***
2050 Corr = 0.086***
2040
2040
2040
2030
2030
2030
2020
2020
2020
2010
0
200
MAF
400
2010
0
100
REF
200
2010
500 1000 1500 2000
World
Figure 7 Relation between regional CO2 budgets and year of peaking of CO2 emissions, for two groups of
climate categories (430-530 in green and 530-650 ppm eq in red). Dots are model scenarios. The blue
lines shows the outcome of quantile regressions, at 5 (dotted), 50 (solid) and 95 (dotted) percentiles.
Pearson’s correlation coefficient is reported along with the confidence interval (***=0.01,
**=0.05,*=0.1). The green markers are bigger for improved clarity.
Summing up, this section has shown that regional carbon budget correlate well with measures of mitigation
efforts which either focus on the long term (e.g. cumulative mitigation, time of zero emissions), or on
emission quotas. However, the relation appears to be less robust when other indicators (e.g. year of
peaking emissions) or second best scenarios (e.g. fragmented cooperation, delayed participation) are
considered.
4.2. Correlation with economic mitigation costs
In this section, we examine the relation between regional carbon budgets and economic indicators of
mitigation policies. Political feasibility of legislating climate policies is heavily dependent on the expected
economic impacts of these policies. Although the global costs of climate stabilization policies are often
found to be relatively modest by IAMs, the regional variations can be much larger (Clarke L., K. Jiang, K.
Akimoto, M. Babiker, G. Blanford, K. Fisher-Vanden, J.-C. Hourcade, V. Krey, E. Kriegler, A. Löschel, et al.
s.d.). Since policymakers care about national and regional impacts on economic activities, it is important to
examine the relation between carbon budgets and economic policy costs8. As discussed in Section 2, in
8
It should be remarked that different models express mitigation costs in different metrics. Top down economic IAMs
use GDP or consumption losses. Bottom up IAM express costs in terms of area under the marginal abatement cost
curve, or in terms of additional energy system costs. Here we combine all metrics, with preference to GDP loss for
those models which report more than one metric.
12
most models, the economic costs of decarbonization can be decoupled from regional costs through
mechanisms such as international carbon trading. Although the actual emissions are independent of the
initial allocation of emission permits, the exchange of the latter in the market can yield benefits or costs,
depending on the trading position. In order to account for policy costs which are independent of the
allocation of emission permits and just depend on the actual emission –our variable of interest- we have
adjusted the economic output produced by the models by removing (adding) the revenues (expenditures)
coming from international carbon trading.
Mitigation Costs (%)
Figure 8 shows the relation between netted mitigation costs and regional carbon budgets. The figure shows
that the correlation between budgets and economic costs is weak, and in some cases not significant. This
result can be ascribed to various factors. First, the economic quantification of mitigation is per se uncertain:
models make very different assumptions on the costs development of different technologies and the
implications of using more expensive technologies for the economy as a whole. Second, if policies are not
‘first best’, costs will depend on the policy structure. However, even when plotting the same chart focusing
only on first best policies, a similar relation is observed. Third, mitigation costs are discounted using a given
net present value, which puts more value on immediate rather than deferred costs. However, even when
looking at different discount rates things do not change much. Finally, an additional key factor for
mitigation costs is the carbon intensity of the economy in the BAU, as well as terms of trade effects for
fossil exporting countries (Stern et al. 2012; Tavoni et al. 2014). This information is not accounted for by the
carbon budgets and therefore it is not surprising to find that there is considerable uncertainty between
regional carbon budgets and mitigation costs.
10
10
Corr = 0.051
5
0
5
0
500
1000
0
Mitigation Costs (%)
Corr = -0.35***
0
5
0
500
1000
Corr = -0.18***
5
0
200
MAF
400
0
-100
ASIA
10
Corr = 0.0086
5
OECD90
10
10
Corr = -0.33***
0
-100 0
0
100 200
LAM
10
Corr = -0.4***
5
100 200 300
REF
0
1000
2000
World
Figure 8. Relation between regional emission budgets and mitigation costs (NPV at 5% discounting), for
two groups of climate categories (430-530 in green and 530-650 ppm eq in red). Dots are model
scenarios. The blue lines show the outcome of quantile regressions, at 5 (dotted), 50 (solid) and 95
13
(dotted) percentiles. Pearson’s correlation coefficient is reported along with the confidence interval
(***=0.01, **=0.05,*=0.1).
Figure S3 shows that a somewhat stronger relation can be established between regional carbon budgets
and the marginal costs of mitigation, e.g. carbon prices (actualized in net present values). But even in this
case, the unexplained variation remains large, testifying to the fact that carbon budgets alone cannot be
used to predict the economic consequences of climate policies. Richer statistical models –including
additional covariates such as relative mitigation effort, carbon intensity etc- can help explain a larger
fraction of the variance of regional economic costs of emission reduction, though both model and regional
specific factors remain important factors (Tavoni & Tol 2010; Stern et al. 2012; Tavoni et al. 2013).
5. Conclusions and recommendations
This paper has assessed the validity and usefulness of regional cumulative emission and carbon budgets to
inform climate policy. The regional focus of the paper is motivated by the policy relevance and importance
of regional policy indicators, such as in the context of the ongoing UFCCC negotiations. Defining the right
metrics of comparability of effort is a key step to evaluate countries climate change mitigation effort (Aldy
& Pizer 2014). Although cumulative emission metrics have not been used in actual policy proposals, they
share the advantages of global carbon budgets, by providing a longer term and more informative signal
than emission quotas in a given year (not to mention less well defined targets). In order to perform this
evaluation, we have used the largest database of scenarios generated by energy-economy-climate models,
the one prepared for the IPCC WGIII 5th assessment report.
Our results suggest an important albeit confined role for regional carbon budgets in climate policy. Thanks
to the linearity between budgets and temperature increase, regional cumulative emissions can be used to
explore the regional contribution to global warming for BAU scenarios. Similarly for the global carbon
budgets, the main limitation is the missing warming contribution of the non-CO2 forcing, which is expected
to be substantial. Our analysis has shown that budgets are also good predictors of mitigation effort, mostly
when this is measured in the long term. The correlation is weaker for shorter term, imperfect, and yet more
widely used metrics such as emissions reductions in a given year or time of peaking or of negative
emissions. Finally, budgets are relatively poor predictors of the economic costs of mitigation. However, it is
difficult to devise single indicators which are effective in forecasting mitigation costs, so this criticism is
applicable to other indicators in general.
Making progress on international climate policy requires comprehensive effort by all the major emitters. In
this sense, developing and testing a variety of indicators of effort is an important area where research can
fruitfully contribute to policy. Regional carbon budgets provide an important step in this direction. More
research is needed to validate and increase the confidence of the regional measures, and to expand the
analysis from the large regional aggregates described in this paper to the country level. We leave these
unanswered questions for future research.
Acknowledgements
14
The authors are grateful for the colleagues in the field of integrated assessment modelling that provided
the model results included in the AR5 scenario database. The research presented has been supported by
the funding of the European Commission DG Research for the PATHWAYS project.
References
Aldy, J.E. & Pizer, W.A., 2014. Comparability of Effort in International Climate Policy Architecture, Rochester,
NY: Social Science Research Network. Available at: http://papers.ssrn.com/abstract=2449645
[Consultato aprile 29, 2015].
Allen, M.R. et al., 2009. Warming caused by cumulative carbon emissions towards the trillionth tonne.
Nature, 458(7242), pagg.1163–1166.
Anderson, K., Bows, A. & Mander, S., 2008. From long-term targets to cumulative emission pathways:
Reframing UK climate policy. Energy Policy, 36(10), pagg.3714–3722.
Azar, C. et al., 2010. The feasibility of low CO2 concentration targets and the role of bio-energy with carbon
capture and storage (BECCS). Climatic Change, 100(1), pagg.195–202.
Blanford, G.J., Rose, S.K. & Tavoni, M., 2012. Baseline Projections of Energy and Emissions in Asia. Energy
Economics. Available at: http://www.sciencedirect.com/science/article/pii/S0140988312001764
[Consultato novembre 8, 2012].
Ciscar, J.-C. et al., A comparability analysis of global burden sharing GHG reduction scenarios. Energy Policy.
Available at: http://www.sciencedirect.com/science/article/pii/S0301421512009251 [Consultato
gennaio 27, 2013].
Clarke L., K. Jiang, K. Akimoto, M. Babiker, G. Blanford, K. Fisher-Vanden, J.-C. Hourcade, V. Krey, E. Kriegler,
A. Löschel, et al., Assessing Transformation Pathways. In: Climate Change 2014: Mitigation of
Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change.
Edmonds, J. et al., 2013. Can radiative forcing be limited to 2.6 Wm−2 without negative emissions from
bioenergy AND CO2 capture and storage? Climatic Change, 118(1), pagg.29–43.
den Elzen, M. & Höhne, N., 2008. Reductions of greenhouse gas emissions in Annex I and non-Annex I
countries for meeting concentration stabilisation targets. Climatic Change, 91(3-4), pagg.249–274.
Elzen, M.G.J. et al., 2012. Reduction targets and abatement costs of developing countries resulting from
global and developed countries’ reduction targets by 2050. Mitigation and Adaptation Strategies
for Global Change, 18(4), pagg.491–512.
Frame, D.J., Macey, A.H. & Allen, M.R., 2014. Cumulative emissions and climate policy. Nature Geoscience,
7(10), pagg.692–693.
Friedlingstein, P. et al., 2014. Persistent growth of CO2 emissions and implications for reaching climate
targets. Nature Geoscience, 7(10), pagg.709–715.
Fuss, S. et al., 2014. Betting on negative emissions. Nature Climate Change, 4(10), pagg.850–853.
15
Gignac, R. & Matthews, H.D., 2015. Allocating a 2 °C cumulative carbon budget to countries. Environmental
Research Letters, 10(7), pag.075004.
Hof, A.F., den Elzen, M.G.J. & Van Vuuren, D.P., 2009. Environmental effectiveness and economic
consequences of fragmented versus universal regimes: what can we learn from model studies?
International Environmental Agreements: Politics, Law and Economics, 9(1), pagg.39–62.
Jacoby, H.D. et al., 2009. Sharing the Burden of GHG Reductions. In J. E. Aldy & R. N. Stavins, a c. di PostKyoto International Climate Policy: Implementing Architectures for Agreement. Cambridge
University Press. Available at:
http://belfercenter.ksg.harvard.edu/publication/18613/sharing_the_burden_of_ghg_reductions.ht
ml [Consultato dicembre 5, 2011].
Kober, T., van der Zwaan, B.C. & Rösler, H., 2013. Emission Certificate Trade and Costs under Regional
Burden-Sharing Regimes for a 2 C Climate Change Control Target. Climate Change Economics.
Kriegler, E. et al., 2013. Is atmospheric carbon dioxide removal a game changer for climate change
mitigation? Climatic Change, 118(1), pagg.45–57.
Kriegler, E. et al., 2014. Making or breaking climate targets: The AMPERE study on staged accession
scenarios for climate policy. Technological Forecasting and Social Change [accepted for
publication].
Luderer, G. et al., 2012. On the regional distribution of mitigation costs in a global cap-and-trade regime.
Climatic Change, 114(1), pagg.1–20.
Matthews, H.D. et al., 2014. National contributions to observed global warming. Environmental Research
Letters, 9(1), pag.014010.
Matthews, H.D. et al., 2009. The proportionality of global warming to cumulative carbon emissions. Nature,
459(7248), pagg.829–832.
Meinshausen, M. et al., 2009. Greenhouse-gas emission targets for limiting global warming to
2[thinsp][deg]C. Nature, 458(7242), pagg.1158–1162.
Miketa, A. & Schrattenholzer, L., 2006. Equity implications of two burden-sharing rules for stabilizing
greenhouse-gas concentrations. Energy Policy, 34(7), pagg.877–891.
National Academy Press, 2011. Climate Stabilization Targets: Emissions, Concentrations, and Impacts over
Decades to Millennia. Available at:
http://www.nap.edu/openbook.php?record_id=12877&page=100 [Consultato settembre 9, 2015].
Raupach, M.R. et al., 2014. Sharing a quota on cumulative carbon emissions. Nature Climate Change, 4(10),
pagg.873–879.
Smith, L.J. & Torn, M.S., 2013. Ecological limits to terrestrial biological carbon dioxide removal. Climatic
Change, 118(1), pagg.89–103.
Stern, D.I., Pezzey, J.C.V. & Lambie, N.R., 2012. Where in the world is it cheapest to cut carbon emissions?*.
Australian Journal of Agricultural and Resource Economics, 56(3), pagg.315–331.
Tavoni, M. et al., 2014. Post-2020 climate agreements in the major economies assessed in the light of
global models. Nature Climate Change, advance online publication. Available at:
16
http://www.nature.com/nclimate/journal/vaop/ncurrent/full/nclimate2475.html [Consultato
gennaio 20, 2015].
Tavoni, M. et al., 2013. The distribution of the major economies’ effort in the Durban platform scenarios.
Climate Change Economics, 4(4), pag.1340009.
Tavoni, M. & Socolow, R., 2013. Modeling meets science and technology: an introduction to a special issue
on negative emissions. Climatic Change, 118(1), pagg.1–14.
Tavoni, M. & Tol, R.S.J., 2010. Counting only the hits? The risk of underestimating the costs of stringent
climate policy. Climatic change, 100(3), pagg.769–778.
Vuuren, D.P. van et al., 2013. The role of negative CO2 emissions for reaching 2 °C—insights from
integrated assessment modelling. Climatic Change, 118(1), pagg.15–27.
van Vuuren, D. et al., 2011. A special issue on the RCPs. Climatic Change, 109, pagg.1–4.
Wara, M., 2007. Is the global carbon market working? Nature, 445(7128), pagg.595–596.
Zickfeld, K. et al., 2009. Setting cumulative emissions targets to reduce the risk of dangerous climate
change. Proceedings of the National Academy of Sciences, 106(38), pagg.16129 –16134.
17
Supporting Online Material
430-530 ppme
530-650 ppme
300
300
2100
100
100
-300
ASIA
REF
OECD90
-300
MAF
-200
LAM
-200
ASIA
-100
OECD90
-100
REF
0
MAF
0
LAM
GtCO2
200
GtCO2
2002050
Cumulative mitigation (%)
Cumulative mitigation (%)
Figure S1 Boxplot of regional traded CO2 budgets (positive=selling, negative=buying), for the periods
2010-2050 and 2050-2100.
100
100
Corr = -0.67***
50
50
0
200
400
600
Corr = -0.52***
0
200
50
0
400
600
100
MAF
200
0
0
100
200
LAM
100
Corr = -0.59***
50
0
0
ASIA
100
Corr = -0.58***
50
OECD90
100
100
Corr = -0.69***
Corr = -0.83***
50
0
100
REF
200
0
500 1000 1500 2000
World
Figure S2: Same as Figure 4 but with carbon budgets and cumulative mitigation to 2050.
18
Cumulative mitigation (%)
Cumulative mitigation (%)
120
100
120
Corr = -0.61***
100
120
Corr = -0.75***
100
80
80
80
60
60
60
40
40
40
0
500
1000
0
500
OECD90
-100
ASIA
120
100
1000
100
100
80
80
60
60
60
40
40
40
400
100 200
120
Corr = -0.83***
80
200
0
LAM
120
Corr = -0.82***
0
Corr = -0.93***
-100 0
MAF
Corr = -0.82***
100 200 300
1000
REF
2000
World
Figure S3: Same as Figure 4 but focusing on the subset of scenarios with delayed or fragmented
participation.
Mitigation (%)
100
100
Corr = -0.67***
50
50
50
0
0
0
-50
-50
-50
200
400
600
400
OECD90
100
Mitigation (%)
100
Corr = -0.57***
Corr = -0.43***
600
0
ASIA
100
100
Corr = -0.41***
50
50
0
0
0
-50
-50
-50
100
MAF
200
0
100
REF
100
200
LAM
50
0
Corr = -0.095***
200
Corr = -0.68***
500 1000 1500 2000
World
Figure S4: Same as Figure 5 but focusing on the subset of scenarios with delayed or fragmented
participation.
19
Carbon price ($/tCO2)
100
100
Corr = -0.33***
50
0
50
0
500
1000
0
Carbon price ($/tCO2)
Corr = -0.25***
0
50
0
500
1000
Corr = -0.053***
50
0
200
MAF
400
0
-100
ASIA
100
Corr = -0.021***
50
OECD90
100
100
Corr = -0.34***
0
-100 0
0
100 200
LAM
100
Corr = -0.29***
50
100 200 300
REF
0
1000
2000
World
Figure S5: Relation between regional emission budgets and carbon prices (NPV at 5% discounting), for
two groups of climate categories (430-530 and 530-650 ppm eq). The blue lines shows the quantile
regressions, at 10, 50 and 90 percentiles.
20

Similar documents