Energy Economics 56 (2016) 564–574

Contents lists available at ScienceDirect

Energy Economics
journal homepage: www.elsevier.com/locate/eneco

Climate policy scenarios in Brazil: A multi-model comparison for energy

André F.P. Lucena a,⁎, Leon Clarke b, Roberto Schaeffer a, Alexandre Szklo a, Pedro R.R. Rochedo a,
Larissa P.P. Nogueira a, Kathryn Daenzer c, Angelo Gurgel d, Alban Kitous e, Tom Kober f
a
Energy Planning Program, Universidade Federal do Rio de Janeiro, Brazil
b
Pacific Northwest National Laboratories, Joint Global Change Research Institute, College Park, MD, USA
c
Pennsylvania State University, College of Agricultural Sciences, University Park, PA, USA
d
São Paulo School of Economics, Fundação Getúlio Vargas (EESP/FGV), Brazil/MIT Joint Program on the Science and Policy of Climate Change, USA
e
European Commission, Joint Research Centre, Sevilla, Spain (The views expressed are purely those of the author and may not in any circumstances be regarded as stating an official position of the
European Commission)
f
Energy research Centre of the Netherlands, Policy Studies, Amsterdam, The Netherlands

a r t i c l e i n f o a b s t r a c t

Article history: This paper assesses the effects of market-based mechanisms and carbon emission restrictions on the Brazilian en-
Received 1 September 2014 ergy system by comparing the results of six different energy-economic or integrated assessment models under
Received in revised form 29 September 2014 different scenarios for carbon taxes and abatement targets up to 2050. Results show an increase over time in
Accepted 6 February 2015
emissions in the baseline scenarios due, largely, to higher penetration of natural gas and coal. Climate policy sce-
Available online 14 February 2015
narios, however, indicate that such a pathway can be avoided. While taxes up to 32 US$/tCO2e do not significantly
JEL classification:
reduce emissions, higher taxes (from 50 US$/tCO2e in 2020 to 162US$/tCO2e in 2050) induce average emission
H23 reductions around 60% when compared to the baseline. Emission constraint scenarios yield even lower reduc-
C88 tions in most models. Emission reductions are mostly due to lower energy consumption, increased penetration
Q40 of renewable energy (especially biomass and wind) and of carbon capture and storage technologies for fossil
Q54 and/or biomass fuels. This paper also provides a discussion of specific issues related to mitigation alternatives
C61 in Brazil. The range of mitigation options resulting from the model runs generally falls within the limits found
C63 for specific energy sources in the country, although infrastructure investments and technology improvements
C68
are needed for the projected mitigation scenarios to achieve actual feasibility.
O57
© 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
Keywords: (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Climate policy
Low-carbon energy scenarios
Mitigation alternatives
Brazil

1. Introduction Most of Brazil's deforestation takes place in the Brazilian

The increase in greenhouse gas (GHG) emissions in the recent de-
cade has been dominated by the emerging economies, explained mainly
by the growth in their economic activity (Peters et al., 2012). In the case
of Brazil, emissions up to 2010 have been dominated by land-use carbon
dioxide (CO2) and non-CO2 gases, pinpointing the key role played by de-
forestation and agriculture in the country and placing it in fourth-place
when it comes to ranking national contributions to observed global
warming (Matthews et al., 2014). When accounting only for CO2 emis-
sions from fossil-fuel burning, cement production and gas flaring, how-
ever, Brazil is ranked as fifteenth (Boden et al., 2013).
⁎ Corresponding author.
E-mail addresses: andrelucena@ppe.ufrj.br (A.F.P. Lucena), larissa@ppe.ufrj.br
(L.P.P. Nogueira).
Amazon, where its rate has decreased substantially in the recent years
(from a 10-year deforestation average of 19,500 km2 year−1 in 2005
to 5843 km2 year− 1 in 2013 — Nepstad et al., 2014). According to
Aguiar et al. (2012), the reduction in deforestation rates in that biome
alone leads to a drop in annual CO2 emissions from more than
1.1 billion tons of CO2 in 2004 to 298 million tons of CO2 in 2011, assum-
ing a direct conversion of lost biomass into carbon. Should deforestation
stabilize at this new level, the energy sector will, in the near future, be-
come the main source of emissions in Brazil.
Globally, Brazil is in a favorable position when it comes to the use of
renewable energy sources. In 2013, over 40% of all primary energy pro-
duced in the country came from renewable sources (EPE, 2014), a value
that is relatively high compared to the world average of around 13%
(IEA, 2013). Most of the renewable sources used in the country come
from sugarcane products (16.1%), hydropower (12.5%) and other bio-
mass (8.3%). Wind, solar and other renewable resources still play a

http://dx.doi.org/10.1016/j.eneco.2015.02.005
0140-9883/© 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 A.F.P. Lucena et al. / Energy Economics 56 (2016) 564–574
small role, with less than 5% of the total primary energy produced in
Brazil (EPE, 2014).
However, socioeconomic development of the country will result in
higher energy use that is not guaranteed to come from renewable sources.
In spite of the current high share of renewables in the Brazilian energy
mix, the country faces a situation where, on the one hand, it needs to in-
crease its energy production to foster socioeconomic development, job
creation and poverty alleviation. On the other hand, the country faces
the near exhaustion of its environmentally feasible hydropower potential
and is expected to increase fossil energy use, with the recent oil discover-
ies in the pre-salt1 fields and the perspectives for increased coal-fired
power generation (EPE, 2013; Nogueira et al., 2014; Saraiva et al., 2014).
Different policy options are available to foster a low-carbon econo-
my. The evaluation of market based policies, such as a carbon tax or ne-
gotiable emission permits, has been widely conducted in worldwide
and regional analyses (Clarke et al., 2012; GEA, 2012; IPCC, 2014). To
date no study has analyzed the effects of different carbon policies,
such as taxes and/or caps, on the Brazilian energy system by running
and comparing different integrated assessment models (IAM).
As part of the Latin American Modeling Project and
Integrated CLimate Modelling And CAPacity building in Latin
America (LAMP–CLIMACAP — van der Zwaan et al., 2016a–in this
issue), six teams have generated profiles of the Brazilian energy sys-
tem out to 2050 under different carbon tax and abatement target re-
gimes using different IAMs. This paper compares the models’ results2
for Brazil in order to assess the possible effects of GHG mitigation
strategies on the country’s energy system. Based on the identifica-
tion of key energy segments provided by this analysis, this paper
provides a discussion of issues particularly relevant to Brazil.
This paper is organized as follows. Section 2 discusses the models
and scenarios used in the study. Section 3 presents the basic assump-
tions and baseline results. Section 4 shows the results for climate policy
scenarios. Section 5 discusses specific issues in the Brazilian energy sys-
tem and relates them to climate change mitigation policies. Finally,
Section 6 concludes the paper with some final remarks.
2. Participating models and scenario description
Within LAMP–CLIMACAP, six modeling teams assessed Brazil as an
independent region and were, therefore, considered in this study.
These groups have produced five scenarios for the Brazilian energy
mix out to 2050 under different climate policy regimes. The models
used in this study are: EPPA (Paltsev et al., 2005, 2013); GCAM (Calvin
et al., 2011); MESSAGE-Brazil (IAEA, 2006; Lucena et al., 2010;
Nogueira et al., 2014); Phoenix (Wing et al., 2011); POLES (Griffin
et al., 2014; Kitous, 2010); and TIAM-ECN (Kober et al., 2014 and van
der Zwaan, 2013, 2016b–in this issue). These models differ from each
other in terms of their modeling approach (optimization or simulation),
spatial resolution (national or global), sectoral scope (partial or general
equilibrium), degree of foresight (myopic or perfect foresight) and rep-
resentation of technological options (type, availability and costs). The
models also differ in how they treat the potential for energy resources
and represent technological change. A comparison of model features
can be found in van der Zwaan (2016b–in this issue) and Clarke et al.
(2016–in this issue).
A baseline scenario and four climate policy scenarios developed
within the LAMP–CLIMACAP exercise are used in this paper and other
studies within this special issue (Clarke et al.; van der Swaan et al.; Cal-
vin et al.; van Ruijven et al.). The current climate policy in Brazil is lim-
ited to 2020 and there is not a clear picture or deep discussion in the
country about a climate policy strategy beyond 2020. Considering this
1
The pre-salt oil fields are so called because of the 2000 m layer of salt above the oil. Es-
timated reserves in these fields range from 30 to 100 billion barrels of oil (OCD, 2009).
2
The results database of the LAMP-CLIMACAP project can be found at https://secure.
iiasa.ac.at/web-apps/ene/LAMPDB/.
565
absence of discussions about possible future mitigation policy choices,
testing standard mitigation instruments, such as carbon prices and
emission targets, can provide useful information for climate policy mak-
ing in the country, given that the model scenarios analyzed here for
these instruments provide cost-effective mitigation options. Additional-
ly, by using a standardized set of policy scenarios, it is possible to com-
pare the effects of these policies across countries in Latin America (e.g.
Clarke et al., 2016–in this issue).
The core baseline scenario is based on business-as-usual assump-
tions at the regional and global levels and is used as the reference for
the climate policy scenarios. It does not include the Brazilian Copenha-
gen Pledge3 or new climate or energy policies except those implement-
ed prior to 2010. The four climate policy scenarios are divided into two
different sets: two scenarios with CO2 price paths applied to all GHGs —
Low CO2 price and High CO2 price; and two others with emission reduc-
tions applied to all fossil fuel and industrial (FF&I) CO2 emissions — 20%
abatement (FF&I) and 50% abatement (FF&I). The scenarios are progres-
sively stringent in terms of mitigation efforts. Both sets of policies begin
in 2020 and all other assumptions are the same as in the baseline.
Table 1, shows the CO2 price paths and emission reductions assumed
by the climate policy scenarios. For a more detailed description of the
scenarios used in this study see van der Zwaan et al. (2016a–in this
issue), van Ruijven et al. (2016–in this issue) and Clarke et al. (2016–
in this issue).
In this paper the results of the different models/scenarios are ana-
lyzed only for the industrial and energy sectors. For an analysis of land
use and forestry emissions resulting from the LAMP–CLIMACAP model-
ing efforts see Calvin et al. (2016-in this issue). Because the sectoral
scope of the models is different – e.g. some models have endogenous
land use modules – in the 20% abatement (FF&I) and 50% abatement
(FF&I) emission reductions are applied to energy and industry only.
Still, not all models include emissions from industrial processes. Howev-
er, since these are relatively small compared to energy emissions, they
do not significantly affect the model comparison (for simplicity, hence-
forth the term ‘emissions from energy’ will refer to emissions from en-
ergy and industry).
3. Basic assumptions and baseline scenarios
Model Projections are largely dependent on the basic assumptions
guiding the evolution of the main drivers for energy production and
consumption. Assumptions about technological development, costs, be-
havior, and trade, vary greatly across models (for more information on
the technological specifications within the models used in the LAMP ex-
ercise, see van der Zwaan et al., in this issue). The models were not har-
monized for Gross Domestic Product (GDP) and population growth,
which creates a broad range of future pathways. The basic population
and GDP assumptions used in each model are described in detail in
van Ruijven et al. (2016–in this issue). Some models are computable
general equilibrium (CGE) models (Phoenix and EPPA) with endoge-
nous GDP pathways that vary across the different scenarios prepared
for this comparison exercise. In all other models, GDP is exogenous
and is the same across scenarios.
The models generally assume that Brazil's population stabilize at dif-
ferent levels by 2050 (except GCAM) and, in some cases, population
peaks in 2040 and then decreases. The assumptions for GDP vary greatly
across models, ranging from a 2.5 to more than a 4-fold increase from
2010 to 2050. In per capita terms, the spread of GDP assumptions is
also large, nearly doubling in the lower case and increasing by 3.6
times by 2050, when compared to 2010.
3
The Brazilian pledges were announced at the UNFCCC Conference of Parties in Copen-
hagen, 2009. These voluntary pledges set emission reduction targets of 36.1–38.9% com-
pared to baseline emissions projected up to 2020 (Brasil, 2009). The extent to which
these pledges are based on a realistic baseline is debatable (see, for example, Lucon
et al., 2013; Clarke et al., 2016–in this issue).
566 A.F.P. Lucena et al. / Energy Economics 56 (2016) 564–574

Table 1
CO2 price paths and emission reductions assumed by the climate policy scenarios.

Scenario Scenario description

Core baseline Business-as-usual scenario including climate and energy policies enacted prior to 2010
Low CO2 price A carbon tax is levied of 10 $/tCO2e in 2020, growing at 4%/year to reach 32 $/tCO2e in 2050.
High CO2 price A carbon tax is levied of 50 $/tCO2e in 2020, growing at 4%/year to reach 162 $/tCO2e in 2050.
20% abatement (FF&I) Fossil fuel and industrial CO2 emissions are reduced by 5% in 2020, linearly increasing to 20% in 2050, with regard to 2010.
50% abatement (FF&I) Fossil fuel and industrial CO2 emissions are reduced by 12.5% in 2020, linearly increasing to 50% in 2050, with regard to 2010.

The wide range in GDP and population growth, however, is not total- and produce biofuels, resulting in lower levels in 2010 compared to the
ly reflected in projected primary energy consumption since different models which include all biomass primary energy. For power generation,
models assume different energy efficiency improvement rates. In all models show a more homogenous composition of the fuel mix for 2010,
models, primary energy roughly doubles from 2010 until 2050 (ranging with little differences across them. For a discussion about base year varia-
from an increase 2.14 to 2.51 times the 2010 levels). As a result, primary tions across models, see van Ruijven et al. (2016–in this issue).
energy intensity decreases in all models, from just below 2010 values to In the baseline scenario the share of fossil fuel increases in all models
almost half of that. In per capita terms, primary energy consumption but POLES due to a large penetration of natural gas and/or coal. In abso-
roughly doubles in all models by mid-century. The reasons for the de- lute terms, these two energy carriers are projected to increase in all
creasing energy intensities vary across models, but generally are related models. Models agree that oil will remain an important energy carrier
to both projected energy efficiency improvements and changes in the in the future, but generally project it to increase at a rate below that of
economic structure of the country. total primary energy. In four out of six models oil consumption in-
The composition of the primary energy mix up to 2050 in the baseline creases little from 2020, despite the large, recently discovered oil re-
scenario is shown in Fig. 1. Some discrepancies can be noticed in 2010, serves in the pre-salt fields.
which are the result of differences in variable definitions, information In all models renewable energy sources increase in the baseline sce-
sources and base year across models. For example, results for biomass in nario. However, their share in total primary energy consumption is
Phoenix only include primary energy that is used to generate electricity projected to decrease by 2050 given the large penetration of fossil

35
30
25
EJ/year

20
15
10
5
0
EPPA

GCAM

MESSAGE-Brazil

POLES

Phoenix

TIAM-ECN

EPPA

GCAM

MESSAGE-Brazil

POLES

Phoenix

TIAM-ECN

EPPA

GCAM

MESSAGE-Brazil

POLES

Phoenix

TIAM-ECN

2010 2030 2050

Oil Coal Gas Nuclear Hydro Biomass Solar Wind Other
1800
1600
1400
TWh/year

1200
1000
800
600
400
200
0
EPPA

GCAM

MESSAGE-Brazil

POLES

Phoenix

TIAM-ECN

EPPA

GCAM

MESSAGE-Brazil

POLES

Phoenix

TIAM-ECN

EPPA

GCAM

MESSAGE-Brazil

POLES

Phoenix

TIAM-ECN

2010 2030 2050

Hydro Oil Coal Gas Nuclear Biomass Solar Wind Other

Fig. 1. Primary energy (upper panel) and electricity (lower panel) mix in the Core baseline scenario for Brazil.
 A.F.P. Lucena et al. / Energy Economics 56 (2016) 564–574
Fig. 2. Emissions from energy in all scenarios for Brazil (Core baseline range shown in all plots, CO2 price scenarios shown in upper row and emission restriction scenarios shown in lower
row).
fuels. This result reflects the expected lower projected cost of fossil fuels
relative to that of renewable energy in a scenario without any kind of
climate policy. The only exception is the baseline scenario from POLES,
which shows a higher penetration of renewable energy sources.
Electricity generation more than doubles by 2050 in all models in the
baseline scenario, ranging from a 2.22 to a 3.18-fold increase. Per capita
electricity consumption increases in all models. In terms of power gen-
eration mix, in the baseline scenario all models projects a diversification
by 2050 compared to today's hydropower-based electricity mix, despite
a still expected increase in hydropower. Most models see a higher share
of fossil-based generation with an increase in natural gas after 2020 and
an increase in coal based generation from 2030 onwards. After 2030–40,
wind and solar power increase in all models (except solar in TIAM-
ECN), but only in a few models do these sources reach a more relevant
share by 2050. Biomass-based electricity generation increases in all
models except in MESSAGE-Brazil, but does not increase beyond a 15%
share of total generation. All models project a small increase in nuclear
energy in the baseline.
Emissions from energy increase in all models in the baselines. Up to
2030, there is a small range across models, which broadens to 2050,
when model results range from 910 to 1665 Mt CO2/year. The range of
fossil fuel and industrial emissions in the baseline scenarios is shown
in Fig. 3; for model specific results we refer to van Ruijven et al.
(2016–in this issue). The results for emissions and GHG mitigation op-
tions are discussed in the next section.
4. Climate policy scenarios
Emissions from energy use for all models are shown in Fig. 2 accord-
ing to different mitigation scenarios vis-a-vis the baseline scenarios. As
expected, the range of results across models for the cap scenarios – 20%
abatement (FF&I) and 50% abatement (FF&I) – is much smaller than for
the tax scenarios — Low CO2 price and High CO2 price. The difference
across the cap scenarios reflects, on the one hand, differences in base
year values and, on the other, differences due to trade between Latin
American countries. These emission reduction scenarios impose restric-
tions across Latin America, thus yielding slightly different results for
Brazil specifically.4 The difference between the two tax scenarios and
4
By being the only country specific model, MESSAGE-Brazil shows lower emissions
than the other models, which implies that emission reduction efforts in the rest of Latin
America are greater than those in Brazil.
567
the baseline scenarios shows evidence that a low tax does not substan-
tially affect emissions from energy and industry in Brazil (average re-
duction from baseline projections around 20% by 2050). A high tax, on
the other hand, induces larger emission reductions in all models to
levels much below the lower boundary of the baseline scenarios (aver-
age reduction from baseline around 60% by 2050). In general, however,
the tax scenarios in this exercise yield less stringent emission reductions
than cap scenarios, the only exception being GCAM.
Emission reductions are induced by a combination of actions, such as
reduced energy demand, decarbonization of primary energy and of
electricity supply mix, carbon dioxide capture and storage (CCS) with
fossil fuels as well as CCS with biomass (BioCCS). The solution found
by different models in terms of the best mitigation alternatives in differ-
ent climate policy scenarios is discussed below.
As mitigation efforts become more intense all models see a reduction
in final energy consumption (Fig. 3). While a low tax does not have
much effect on final energy consumption, in more stringent scenarios
some models project reductions in energy consumption that are half
of the baseline levels by 2050, which is very close to 2010 levels
(EPPA and Phoenix).5 Other models see much lower reductions (e.g.
MESSAGE-Brazil, GCAM and POLES at around 10% below baseline levels
in very stringent scenarios). The same result applies in terms of primary
energy, except for MESSAGE-Brazil. In all models, conversion efficiency
from primary to final energy decreases as mitigation efforts change the
primary energy mix towards less-efficient-to-use energy sources, such
as biomass. In MESSAGE-Brazil, this effect is large enough to make pri-
mary energy consumption increase above baseline, even while final en-
ergy consumption decreases.
Besides reducing energy consumption, mitigation also occurs
through changes in the primary energy mix of the country, which is
shown in Fig. 3. In general, a low carbon tax does not affect the primary
energy mix by a large extent. In the higher tax or emission restriction
5
These are CGE models, which consider the impacts of the policy on the overall econo-
my. As so, there are four main channels leading to the decrease in energy consumption:
i) GDP decreases under the carbon policies, reducing the per capita income and, as a con-
sequence, the energy consumption relative to the baseline scenario; ii) carbon policies in-
crease the costs of the fossil energy sources, which reduces consumption; iii) higher
energy prices induce energy efficiency; and iv) as carbon policies are applied to other
countries, there is a decrease in the global economic activity, reducing again the use of
energy.
568 A.F.P. Lucena et al. / Energy Economics 56 (2016) 564–574

40
2030
35

30
EJ/year 25

20

15

10

0
EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN

EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN

EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN

EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN

EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN
Core baseline $10 CO2 price $50 CO2 price 20% abatement (FF&I) 50% abatement (FF&I)
Oil Oil w/CCS Coal Coal w/CCS Gas Gas w/CCS Nuclear Hydro Biomass Bio w/CCS Solar Wind Other

40
2050
35

30

25
EJ/year

20

15

10

0
EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN

EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN

EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN

EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN

EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN
Core baseline $10 CO2 price $50 CO2 price 20% abatement (FF&I) 50% abatement (FF&I)
Oil Oil w/CCS Coal Coal w/CCS Gas Gas w/CCS Nuclear Hydro Biomass Bio w/CCS Solar Wind Other

Fig. 3. Primary energy mix in baseline and climate policy scenarios for Brazil.

scenarios the primary energy mix changes substantially with regard to within the 2050 time horizon. However, most of the power system ex-
the baseline. pansion is based on other sources/technologies. The penetration of fossil
As expected, a decrease in fossil fuel consumption is observed in cli- fuels, which is high in the baseline scenario, greatly decreases in climate
mate policy scenarios. Coal loses importance in some models in re- policy scenarios, being replaced by renewable energy sources and/or
sponse to the low tax but faces large reductions in all models in the converted to facilities coupled with CCS. As GHG mitigation policy be-
other climate policy scenarios. Most of the remaining coal consumption comes more stringent, models see a penetration of wind, biomass and
in stringent climate policy scenarios is used with CCS technologies. Nat- solar, though the mix of these sources varies from model to model. Nu-
ural gas, however, remains relevant in all models and scenarios (except clear energy also increases in all climate policy scenarios.
EPPA), in some cases with CCS. Results show that mitigation policies,
such as carbon taxes or abatement targets, have some effect on oil con- 5. Implications of future pathways to the Brazilian energy industry
sumption. However, results show, in general, that oil consumption does
not decline much as GHG mitigation efforts increase. One reason is the 5.1. Fossil fuels
rapidly increasing demand for transport services and the high costs to
switch from oil products to alternative fuels, which is, in the cases of Although models point to a lower use of fossil fuels in climate policy
electricity and hydrogen, accompanied by considerable infrastructure scenarios, these energy sources remain relevant in the Brazilian energy
investments. mix through 2050 in all scenarios, either coupled with CCS technologies
In terms of renewable energy, the importance of biomass increases or not. Below a discussion about issues related to the use of fossil fuels in
in all models as mitigation policies become more rigorous (except Brazil is made. CCS is discussed in Subsection 5.4.
TIAM-ECN). Some models make use of negative emissions from BioCCS
to achieve climate policy objectives. In some models, biomass becomes 5.1.1. Coal
the major primary energy source in climate policy scenarios by 2050. Under mitigation scenarios, models show that coal would become
Solar and wind increase in climate policy scenarios, but only in TIAM- feasible when equipped with CCS. In a baseline scenario, or under low
ECN and POLES do they reach a relevant share by the end of the period. carbon taxes, expansion of primary energy consumption includes a
Results for the mix of sources/technologies used in electricity gener- high penetration of coal. Current projects underway in Brazil corroborate
ation in 2030 and 2050 are presented in Fig. 4. Hydropower is currently such a scenario. Brazil's current installed coal-based power plant capac-
the major source of electricity generation and should remain important ity includes ten power plants totaling 2 GW which represents roughly
 A.F.P. Lucena et al. / Energy Economics 56 (2016) 564–574 569

2000
1800
2030
1600
1400

TWh/year
1200
1000
800
600
400
200
0
EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN

EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN

EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN

EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN

EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN
Core baseline $10 CO2 price $50 CO2 price 20% abatement (FF&I) 50% abatement (FF&I)
Hydro Oil Coal Coal w/CCS Gas Gas w/CCS Nuclear Biomass Biomass w/CCS Solar Wind Other

2000
1800
2050
1600
1400
TWh/year

1200
1000
800
600
400
200
0
EPPA

MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN

EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN

EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN

EPPA
GCAM
MESSAGE-Brazil
POLES
Phoenix
TIAM-ECN

EPPA
GCAM
MESSAGE-Brazil
POLES
GCAM

Phoenix
TIAM-ECN
Core baseline $10 CO2 price $50 CO2 price 20% abatement (FF&I) 50% abatement (FF&I)
Hydro Oil Coal Coal w/CCS Gas Gas w/CCS Nuclear Biomass Biomass w/CCS Solar Wind Other

Fig. 4. Electricity mix in the baseline and in the climate policy scenarios for Brazil.

1.6% of the country's total installed capacity. This capacity is about to in- 20 million tons per year mostly for the iron and steel industry (EPE,
crease in the near future, as three new coal power plants, totaling some 2014). Expansion in the use of imported coal for power generation
1.4 GW, are currently under construction in the northeast, which will would likely take place in the northeast and southeast regions (EPE,
run on imported coal, mostly from Colombia (ANEEL, 2014). In the 2007). EPE (2007) estimates that major ports are capable, without the
near future, six additional new coal power plants will be built, adding need for additional investments, of receiving an amount of imported
more 3.4 GW to the system (ANEEL, 2014). coal sufficient to supply 10 GW of coal-based power plants with an av-
Coal reserves in Brazil are located mostly in the southern region of erage capacity factor of 0.75 (around 65 TWh/year). Further analysis
the country. While BP (2014) estimates proven reserves (90 percent should evaluate the possibility of expanding this coal import capacity
probability) of coal in Brazil of 7068 million tons, DNPM (2013), using if values projected by some models are to become feasible.
a deterministic approach, estimates measured reserves of about
6710 million tons. However, the quality of Brazilian coal is relatively 5.1.2. Oil and gas
low (average 3200 kcal/kg and high ash content), leading to its use in Although models show that oil consumption in Brazil would be af-
subcritical fluidized-bed-combustion power plants with low first-law fected by climate policy, it remains relevant in all scenarios and models.
efficiencies (around 33%), which yields an estimated theoretical poten- Brazil has offshore petroleum reserves and has recently discovered re-
tial of 28 GW in new coal-based power plants running on domestic coal, sources in the pre-salt fields. While the pre-salt offshore fields are the
with a capacity factor of 0.60 (EPE, 2007). most relevant petroleum exploration frontier in Brazil, there are large
A greater use of domestic coal in thermal power plants has been uncertainties regarding the amount of resources and reserves in
studied by different authors (Antunes, 2009; Gavronski, 2007; this area. Estimates of reserves in the pre-salt vary greatly. Low
Hoffmann, 2010; Ortiz, 2011). They show that, given the low quality estimates consider 30 billion barrels (Costa and Souza-Santos,
of the Brazilian coal, all plants running on domestic coal already are, 2009; Denmark, 2009), while the more optimistic estimates reach
and need to be, located in the south of the country, near the mines. As 100 billion barrels (Denmark, 2009; Fishman, 2010; Maugeri, 2012).
such, the exploration of this potential will require new investments in Saraiva et al. (2014) applied a modified multi-Hubbert model to es-
transmission lines to export the excess of electricity from coal-based timate Brazil's oil production curves according to different ultimate re-
power expansion in the south. coverable resource (URR) scenarios for offshore fields. Assuming a low
High levels of penetration of coal in the Brazilian energy mix, there- estimate for pre-salt reserves (30 billion barrels), oil production in the
fore, would rely on imported coal. Brazil currently imports some country would peak in 2027 at 4.9 million barrels a day (Mb/day),
570 A.F.P. Lucena et al. / Energy Economics 56 (2016) 564–574

while assuming 50 million barrels of reserves (USGS, 2000) would lead 2014), the cash flow of shale exploration is based on regular CAPEX and
to a peak production of 5.4 Mb/day in 2034. In the latest case, the sce- OPEX expenditures given the need to constantly drill new wells (Lage
narios of Saraiva et al. (2014) show the possibility of sustaining an oil et al., 2013). This constant drilling rhythm may be hindered by institu-
production at levels higher than 4 Mb/day for almost 30 years. This fig- tional barriers related to the lack of speed to cope with a high frequency
ure is almost twice the current petroleum production in Brazil.6 of biding rounds, logistic barriers to transport this gas to consumer cen-
Therefore, even without considering the most optimistic estimates ters and the lack of capacity of the national industry to attend such a
of reserves in the country, Brazil would become a major oil producer, high demand for equipment and services.
provided that these resources can, in fact, become reserves and be put
into production. There are many challenges related to exploration and 5.2. Renewable energy
production of oil in the Brazilian pre-salt, such as technical, economic
and institutional barriers. The reserves of the pre-salt cluster are located All models indicate that renewable energy sources will increase their
at a depth of over 6000 m, below a salt layer of about 2000 m, and there role in Brazil's energy system in climate mitigation scenarios, besides
are many challenges to extracting hydrocarbons in such conditions the already and sustained relevant role of hydropower in the country's
(PETROBRAS, 2011). Still, such high levels of production would lead electricity generation. This is increasingly important for scenarios with
the country to a position of a large oil exporter. One issue, that is rele- higher carbon tax or carbon caps. Nevertheless, there remain challenges
vant for Brazil, is the extent to which the country would be able to pro- either in keeping the role of hydropower in the electricity generation or
duce those reserves in a stringent mitigation scenario in which a global in increasing the role of wind, biomass and solar. This section addresses
tax, for instance, would be applied. Country specific models (e.g. these issues.
MESSAGE-Brazil) are limited in this regard since they assume that sur-
plus oil production can be exported. Global models have the advantage
5.2.1. Bioenergy
of capturing global effects of a climate policy and, therefore, the possibil-
The results of the models analyzed in this study point to biomass as a
ity of restrictions in the global market for oil which could, eventually,
major energy source in Brazil up to 2050, both in the baseline scenario
hamper the production of oil in Brazil. The global models analyzed in
and, with increasing importance, as GHG mitigation targets become
this study that reported oil trade project high exports of oil up to 2050
more stringent. This is true in terms of primary energy and electricity
in all scenarios, regardless of global climate policy.
generation. Therefore, modern biomass consumption in Brazil is (and
The model results point to natural gas as a crucial energy source for
would remain) mainly associated with electricity generation and pro-
Brazil through 2050. In most scenarios, natural gas remained a relevant
duction of liquid biofuels. Although charcoal is also relevant, its use is
energy source regardless of climate policy. Brazil has historically been a
limited to the iron and steel industry.
natural gas importer (e.g. from Bolivia) and in all models/scenarios the
As of today, electricity from biomass is derived mostly from
country will continue to be a net importer. Most of the country's pro-
combined-heat-and-power facilities fuelled with sugarcane bagasse,8
duction of natural gas is associated with oil production and this will
which itself is driven by ethanol demand. Thus, biomass is strictly linked
probably continue to be the case given the large offshore oil production
to both electricity and liquid fuel markets in Brazil. Although this has
frontier of the pre-salt basins.
some advantages, an eventual decrease in ethanol demand in the future
On the one hand, onshore natural gas potential production faces
caused by, for instance, a replacement of the current auto-fleet by elec-
challenges related to the expansion of the gas transportation grid,
tric vehicles would, on the one hand, increase electricity demand while,
given its high upfront costs and the monopolistic nature of the
on the other, decrease the availability of biomass (sugarcane) for elec-
Brazilian market (Camargo et al., 2014; Mathias and Szklo, 2007). On
tricity generation. Models used in this study, however, did not see a
the other hand, transporting the production of offshore pre-salt associ-
large penetration of electric vehicles up to 2050.
ated gas to consumer centers is costly, given the far distance from the
The amount of agro-industrial residues that can be converted to
coast (round 300 km). Additionally, associated natural gas in pre-salt ba-
electricity in Brazil is much higher than what is being currently used
sins has a high CO2 content (11–40% on average, volume basis) requiring
(MAPA, 2014). For instance, Portugal-Pereira et al. (2014) sized
separation from the natural gas to allow the fuel's transportation and
Brazil's bioenergy potential according to the theoretical capacity of bio-
commercialization (avoiding clathrate formation). Therefore, CCS (in-
mass production,9 its environmental impacts, and technical–economic
cluding enhanced oil recovery — EOR) in pre-salt petroleum basins is
viability. Overall, their findings indicate that the technical potential of
needed.7 Besides costs, which are still relevant, a main issue is the car-
agricultural and agro-industrial residue conversion to electricity was
bon capture equipment footprint and the restricted area available in
nearly 141 TWh/year in 2010 (compared to the current 38 TWh10 —
ultra-deep water oil platforms. Imperio et al. (2014) estimated that
EPE, 2014). Nearly 88% of the total potential derives from residues of
only one module of carbon capture membranes would fit in platforms
sugarcane, soybean and maize crops, as these are major cash crops in
designed for pre-salt basins in Brazil. This would not be sufficient to pro-
Brazil. In addition, the adoption of better thermodynamic cycles in the
cess all the natural gas produced. Therefore, assuming no improvement
country's current biomass-fuelled thermal-power plants and the in-
in membrane performance, natural gas reinjection will likely increase,
creasing possibility of using dedicated forestry resources to power gen-
which means that less saleable gas will be produced as it will be progres-
eration (Hoffmann and Szklo, In Press) can expand the already large
sively diluted.
potential for electricity production from biomass.
Finally, Brazil holds resources of unconventional gas, but the extent
In general, there are no major challenges related to direct and indi-
to which these resources will have an important role in the country's
rect land use changes which could undermine the models’ results for
energy mix is unclear. Although the level of resources in the country is
electricity from biomass. However, increasing use of biomass,
estimated to be the 10th largest in the world (245 trillion cubic feet —
EIA, 2011), the level of geological knowledge of shale gas resources in 8
Black liquor fuelled cogeneration plants and some small power plants based on char-
Brazil is still very low, making estimates very uncertain. Nevertheless,
coal or elephant grass can also be found in Brazil (ANEEL, 2014).
the challenges that unconventional gas faces in Brazil are huge. Regard- 9
The theoretical capacity defines the maximum available bioenergy under biophysical
less of environmental restrictions and water availability (Camargo et al., and agro-ecological conditions that hold down the growth of crops and residues, such as
temperature, solar radiation, rainfall, and soil properties. This theoretical potential is albeit
limited by environmental constraints, as agricultural residues are important biome regu-
6
According to BP (2014), only four countries currently produce more than 4 Mb/day: lators (Portugal-Pereira et al., 2014). The authors do not include electricity generation
Saudi Arabia, Russia, United States and China. from black liquor in the paper and pulp industry.
7 10
In fact, this has been treated as a baseline scenario by Petrobras, the Brazilian oil Excluding electricity generation from black liquor so as to be comparable to the values
company. estimated by Portugal-Pereira et al. (2014).
 A.F.P. Lucena et al. / Energy Economics 56 (2016) 564–574
particularly through dedicated forestry in specific regions of Brazil, can
have negative local impacts (Hoffmann and Szklo, In Press). The same
is valid for liquid biofuels. According to Leal et al. (2013), yield improve-
ments and the combination first- and second-generation biofuels would
allow an ethanol production of some 300 billion liters in 2030 using
22 Mha of land.11 These numbers are not a cause of concern on a nation-
al scale, but under a local perspective they could cause major impacts on
traditional agriculture (Leal et al., 2013).
The extent to which the use of biomass is effective for mitigating
GHG emissions, however, depends on the life cycle of such energy
sources. Some models used in this study only account for direct emis-
sions, while others have an endogenous land use model. Nevertheless,
there are many uncertainties surrounding the methodologies in use
today for life-cycle analysis of GHG emissions for land use change
(Larson, 2006; Nassar et al., 2011). Calvin et al. (2016-in this issue) ex-
plores this complex issue in more detail in order to evaluate whether or
not land use might influence the amount of mitigation required from
the energy sector.
5.2.2. Hydropower
Hydropower is currently the main source of electricity in Brazil
(around 80%, on average, over the last ten years — EPE, 2014) and is
projected to remain important over the coming decades according to
the models assessed in this study. All models see, in absolute terms,
some increase in hydropower generation up to 2050.
Hydropower expansion in Brazil is controversial. The inventoried re-
maining potential is still large, estimated at 126 GW. However, around
60% of that (including the best sites) is located in the Amazon basin
and another 10% in regions with high environmental impacts elsewhere
in the country (EPE, 2008). There still remain important environmental
concerns and social–politic conflicts associated with the construction of
dams and the consequent flooding of large areas in those regions
(Bermann, 2012; Pimentel, 2012). Thus, increasing the production of
hydropower in Brazil can generate pronounced local environmental
and social impacts.
One implication of the potentially high environmental impacts is
that the share of remaining potential to be eventually exploited will
likely be based on run-of-the-river hydropower plants with smaller res-
ervoirs so as to minimize local environmental impacts.12 These plants
are most vulnerable to climate since river flow can be highly variable,
especially across seasons. Reservoir storage capacity can compensate
for seasonal (or even annual) variations in river flow, enabling electric-
ity generation throughout the year and matching varying power de-
mand. In the operation of the Brazilian interconnected system, thus,
hydropower production based on new run-of-the-river plants would
need to be increasingly complemented by other power sources. Also,
global climate change can add a significant amount of uncertainty to
the climate variability and, hence, to the planning and operation of hy-
dropower (Lucena et al., 2009 and 2010).
5.2.3. Wind power
While models do not project a large penetration of wind in the base-
line scenario (except POLES), wind power generation expands consider-
ably with climate policy. However, all models show a share of wind
below 20% of total generation, possibly because of operational con-
straints, which indicates that the contribution of this source to climate
change mitigation may be limited.
Most of the high-quality wind resources are found in the north-east
and south of Brazil. Distinct sources of information provide different
estimates for the total onshore wind power potential in Brazil. While
11
Embrapa (2009) indicates that 65 Mha of land can be devoted for sugarcane without
significant impacts on food production and on the environment.
12
This has been the case, for example, of the Belo Monte hydropower plant (11 GW) in
the Brazilian Amazon, whose reservoir is now two thirds the size of the original project.
571
CEPEL (2001) estimated, based on 50-meter-high measurements,
some 143 GW (yielding, in average, 272 TWh/year), other estimates,
that assume the deployment of larger wind turbines capturing stronger
wind resources at higher elevations (100-meters and higher) quote
some 300 GW (Simões, 2010), or even 350 GW (GWEC, 2011), in
some regions with average annual capacity factors reaching 0.40
(Borba et al., 2012) or even 0.50 (IEA, 2013). Therefore, the wind
power potential in Brazil is not a constraint for the expansion of this
technology over time in the country. In fact, wind energy has become
increasingly important recently, increasing from nearly zero in 2005
to a currently installed capacity of 3.1 GW. In addition to that, as of
today, plants with 3.3 GW of capacity are under construction and
other plants with 5.9 GW of capacity are being licensed (ANEEL, 2014).
However, large deployments of this technology still face technical
and operational challenges (Borba et al., 2012). Up to a 20% limit “the in-
tegration of wind energy generally poses no insurmountable technical
barriers and is economically manageable” (Wiser et al., 2011, pp. 560).
Nevertheless, the degree of penetration of wind that a system can inte-
grate depends on the characteristics of the system. Brazil has some op-
portunities for integrating wind and hydropower, for example in the
northeast of the country due to seasonal complementarities between
these two sources (Simões, 2010). Interestingly enough, the anticipated
growth in hydro-capacity in Brazil based on run-of-river projects, which
are very much dependent on seasonal variations, can, in fact, with prop-
er electric power interconnections, be balanced, at relatively low inte-
gration costs by the seasonal supply patterns of wind in the northeast
of the country (IEA, 2013). However, the degree to which wind power
can be actually integrated into the Brazilian grid has not yet been thor-
oughly assessed and the models used in this study are not capable of ex-
ploring this issue at full length, which may lead to an underestimation of
the penetration of wind energy in Brazil. The models used here are fo-
cused on long-term system expansion and do not have the ability to
evaluate operational strategies in detail. For that, dispatch models
adapted to the particularities of the Brazilian interconnected system
are needed. Furthermore, looking at the 2050 horizon, power system
operation will depend on the thechnologies used in the expansion of
the Brazilian system until then.
5.2.4. Solar power
Although all models assessed here see some penetration of solar en-
ergy through 2050, only POLES and TIAM-ECN see a large penetration of
this technology in this time frame. Brazil has abundant global solar re-
sources. Solar incidence in the country ranges from 1500 kWh/m2 per
year to 2153 kWh/m2 per year13(SolarGIS, 2013). The region with the
highest average overall daily radiation is the northeast of the country,
a semi-arid region. Nevertheless, currently Brazil only has about
10 MWp of photovoltaic (PV) installed capacity (ANEEL, 2014).
Centralized solar generation participated in the national auctions for
contracting new power in 2013 without success, which indicates that
large-scale centralized plants are not yet cost-competitive under the
current Brazilian electric power market conditions (Miranda et al.,
2014). However, the potential for concentrated solar plants (CSP) is
huge14 and some specific arrangements could improve feasibility in
the medium to long term. For example, Malagueta et al. (2014) show
that the hybridization of CSP plants can reduce the levelized cost by
30% to 50% (depending on the location) for two main reasons: increased
annual production (higher capacity factors) based on the combustion of
natural gas or biomass; and reduction in the solar multiples and the use
13
For comparison purposes, Germany, which is the country with the largest installed
photovoltaic capacity in the world, receives about 1300 kWh/m2 per year (SolarGIS, 2013).
14
Considering restrictions related to, e.g., water and land availability, the annual techni-
cal potential for CSP using parabolic trough plant potential is around 1900 TWh, while the
technical potential of solar towers with thermal storage is around 1000 TWh (Burgi,
2013).
572 A.F.P. Lucena et al. / Energy Economics 56 (2016) 564–574

of fewer collectors. Also, by allowing the combination of solar and bio-
mass residues for power generation, hybridization could help by gener-
ating income in the poor semi-arid regions in the northeast of Brazil.
In the case of PV, although centralized generation is not yet compet-
itive, distributed generation may be an alternative since, from the stand-
point of the final consumer, electricity prices are perceived, rather
than electricity generation costs (Miranda et al., 2014). Some 70% of
the value of electricity tariffs in Brazil account for taxes, transmission
and distribution costs (Fugimoto, 2010). By comparing costs and tariffs
for roof-top installed PV in Brazil, Miranda et al. (2014) show that, from
an economic point of view, around 68,000 households could justify the
installation of photovoltaic arrays on their roofs. However, the imple-
mentation of large scale distributed PV generation would also depend
on the development of PV services in the country.
5.3. Nuclear energy
Although not having a significant role in the scenarios analyzed, a
small expansion of nuclear energy was observed in all models. Some
models/scenarios only include the one nuclear plant currently under
construction (Angra III), whereas extreme scenarios see a four-fold ex-
pansion in nuclear generation by 2050.
The implementation of nuclear plants in Brazil is highly controver-
sial. Actually, there have been many debates in Brazil about the comple-
tion of Brazil's third nuclear power plant, Angra III (Cabrera-Palmer and
Rothwell, 2008), while different government plans suggest an expan-
sion that would add 6 GW up to 2030 (Eletronuclear, 2011; EPE,
2007). On one side, the country's nuclear industry emphasizes the ad-
vantages of this option (including carbon emission mitigation, but fo-
cusing more on technological development and energy security). On
the other side, experts highlight the problems of cost overruns, spent
nuclear fuel management (which is still an issue in other countries,
Singer 2013), and inflexible operation of nuclear plants (Borba et al.,
2012). For example, Carvalho and Sauer (2009) analyzed the cost over-
runs of the Angra III nuclear power plant, with construction lasting for
more than two decades,15 and concluded that, given the country's ener-
gy resources, “the potential decision to finish construction work on
Angra III simply to justify the existing sunk could prove to be a mistake”.
They also contested the opinion of some Brazilian policy-makers and
experts that, given the Brazilian uranium reserves16 and the country's
technical and industrial capability to enrich it (Cabrera-Palmer and
Rothwell, 2008), nuclear plants would be also justifiable by the need
to maintain the country's technological development and expertise.
5.4. CCS and BioCCS
The models identified CCS as a potential technology to decarbonize
the energy sector. However, to date, there are no commercial CCS appli-
cation in the power sector or in energy-intensive industries and only a
handful of large-scale demonstration projects are in operation or under
construction worldwide, with only one located in Brazil (Santos Basin,
see Subsection 5.1.2). In 2012, a pilot plant was designed to demonstrate
CO2 capture in a planned coal-fired thermal power plant in Brazil
(Rochedo and Szklo, 2013a). However, this project was discontinued.
Studies performed for Brazilian coal or gas fired-thermal power plants
identified that mitigation costs hover between 70 and 100 US$/tCO2,
not considering the risk of being pioneer plants (Hoffmann, 2013;
Hoffmann et al., 2012; Rochedo and Szklo, 2013b), which can increase
the costs by more than 50% (Hoffmann and Szklo, 2011). This range
15
Angra III is planned to start operation in 2018.
16
Regardless of the debate, from a long term perspective, uranium supply is not a restric-
tion to the expansion of nuclear energy in Brazil (EPE, 2007). According to a simple esti-
mate, if all of Brazilian uranium reserves (309,370 tU — EPE, 2014) would be used with
an efficiency of 28.5 kg U/GWh, it would be possible fuel of 22.9 GW of nuclear for
60 years.
considers capture and compression, which represent around 70–80%
of the whole carbon capture system cost (Rochedo and Szklo, 2013b).
Carbon transportation and storage, although are not the main cost
drivers, might face a huge barrier, since carbon transportation requires
an institutional arrangement to deal with social acceptability, hub plan-
ning, property rights, tariff definition and storage monitoring (Costa,
2014). Brazil already faces challenges to expand the country's natural
gas grid, as mentioned above. Natural gas is a tradable fuel, while CO2
is a negative externality, for which the value is associated with carbon
mitigation policies.17 Hence, it can be expected that it will be challeng-
ing to establish an institutional arrangement for the construction of CO2
pipelines in Brazil (Costa, 2014).
The CCS potential in Brazilian thermal power plants should also con-
sider that today's expected major technological option (post-combus-
tion capture with amines, according to Rochedo and Szklo, 2013a)
requires additional water consumption and a concurrent generation of
residues. Some studies performed for Brazil indicated that some regions
of the country would not support the higher water demand that carbon
capture systems would eventually add to existing or planned coal fired
plants (Hoffmann et al., 2014; Merschmann et al., 2013).
Moreover, given the large energy penalties of carbon capture sys-
tems, CCS increases the coal consumption of a power plant. Branco
et al. (2013) performed a life-cycle assessment (LCA) for a coal-fired
power plant in Brazil and showed that, with CCS, a plant which captures
90% of its CO2 would have its GHG avoidance potential reduced to 72%
when accounting for indirect emissions. This is the result of an increase
in coal consumption and the associated CH4 emissions at the coal-
mining stage.18
In sum, the implementation of CCS in Brazilian thermal power plants
faces considerable challenges, qualitatively equal to the ones faced glob-
ally, but intensified by the lack of major domestic technological and in-
stitutional development related to this option. However, BioCCS and
even CCS in hydrogen generation plants in petroleum refineries show
brighter prospects. BioCCS in ethanol distilleries presents no major
technical challenges for capturing, given the higher CO2 content in the
fermentation exhaust. The same is valid for hydrogen production plants
with CCS. Some studies also consider the possibility of using CO2 for pro-
ducing chemicals, e.g. methanol, thus avoiding the problem of carbon
transportation and storage (Farias, 2014).
6. Final remarks
This study compared scenarios produced by six modeling teams
with different baseline assumptions for GDP, population, energy costs
and technological development. Although Brazil’s current energy mix
has a relatively low carbon intensity due to a high share of renewable
energy, the baseline scenarios through 2050 suggest that this picture
would radically change. Models project, for the baseline scenario, a 2
to 3.5-fold increase in emissions from energy in relation to 2010 as a re-
sult of a growing penetration of fossil fuels. This indicates that, without
dedicated climate policies, Brazil may take a pathway distinct from that
required by all countries to keep the average rise in global temperature
to below 2 °C, when compared to pre-industrial levels, as indicated by
the Copenhagen Accord of 2009 and the Cancun agreements in 2010
(UNEP, 2013).
Nevertheless, models show that climate policies can reverse this
pathway in Brazil. High carbon taxes may induce large reductions in
emissions when compared to baseline. The emission constraint scenar-
ios, in turn, showed that emission levels below that of 2010 are techno-
logically feasible when considering industry and energy emissions.
Mitigation would be achieved by a combination of actions, such as ener-
gy demand reductions, decarbonization of primary energy and of
17
Although CO2 can have a price when used in EOR, this does not imply that there is a
large market for CO2.
18
Based on a global warming potential metric with a 100-year horizon.
 A.F.P. Lucena et al. / Energy Economics 56 (2016) 564–574
electricity supply mix, fossil-fuel-based CCS and BioCCS. Also, agricul-
ture, forestry and other land use (AFOLU) represent a largely relevant
source of emissions in Brazil. Actions to reduce emissions from AFOLU
can further reduce emissions in Brazil or even reduce the mitigation
burden in the energy sector (see Calvin et al., 2016-in this issue). A dis-
cussion of the Brazilian energy context and resources potential indicates
that the range of mitigation options resulting from the models' runs
generally falls within the limits for specific energy sources in the coun-
try. However, in most cases infrastructure investments and technology
improvements are needed for the projected mitigation scenarios to
achieve actual feasibility.
Based on these results, a few recommendations for energy and
climate policies in Brazil are made. Initially, setting up market-
based instruments, such as those analyzed here, requires aligning
domestic prices of fossil fuels to international ones, enabling fluctu-
ations in the latter to be transferred to final consumers. This is espe-
cially true for oil products, which have a long history of price control
in the country. Also, climate policy should not only be limited to
market-based instruments, such as the ones analyzed in this paper.
Dedicated climate policies should also focus on the opportunities
and barriers for the implementation of mitigation actions in the
country. For example, given that some level of mitigation in the
Brazilian energy sector could be achieved by the use of technologies
not yet technically and/or commercially mature (e.g. CCS and
BioCCS), investments in research, development and demonstration
(RD&D) are important to increase the range of options available in
the future. Finally, mitigation strategies should take into account
the country's energy system vulnerability to climate change impacts
and seek initiatives that foster local/regional adaptability to those
vulnerabilities through adaptation actions and socioeconomic
development.
Long-term scenario building is subject to a large degree of uncertain-
ty. This paper discussed some particular aspects of options that were se-
lected by the simulations of a set of model. These models, in turn, have
their own limitations in terms of the database used, how well they rep-
resent the complexities of the Brazilian energy system and their meth-
odological framework. Furthermore, the results of the models do not
account for market barriers that may hamper the implementation of
mitigation options. For example, coal and nuclear are not currently
adopted in large scale in Brazil and might face some political and even
industrial barriers to be installed. Also, some of the mitigation actions
assessed by the different models are based on technologies that are
not yet fully consolidated, such as fossil-fuel-based CCS and BioCCS. In
this regard, conducting sensitivity analysis to cover some of these un-
certainties, providing, for instance, an analysis of whether stringent mit-
igation targets could be achieved without making use of CCS based
options, is a valuable contribution of future research. Further discussion
as to how the proposed carbon prices could be achieved in practice is
also relevant and should be addressed in future studies. Finally, al-
though not considered in this study, the implications of the projected
high use of biomass on AFOLU emissions are discussed in Calvin et al.
(2016-in this issue).
Acknowledgments
We acknowledge the funding from US EPA and USAID, through the
LAMP, and the European Union, through the CLIMACAP project
(EuropeAid/131944/C/SER/Multi), for the contributions made to the
several modeling teams. We would like to thank the feedback and ef-
forts from all CLIMACAP and LAMP project partners for enabling the re-
search results reported in this article. We acknowledge the financial
support from CNPq and ANP. Finally, we especially thank Régis
Rathmann, Rafael Soria, Mauro Chavez-Rodrigues, Joana Portugal-
Pereira and Alexandre Koberle for their help in some stages of the
LAMP–CLIMACAP exercise.
573
References
Aguiar, A.P.D., Ometto, Jean Pierre, Nobre, Carlos, Lapola, David Montenegro, Almeida,
Claudio, Vieira, Ima Célia, Soares, João Vianei, Alvala, Regina, Saatchi, Sassan,
Valeriano, Dalton, Castilla-Rubio, Juan Carlos, 2012. Modeling the spatial
and temporal heterogeneity of deforestation-driven carbon emissions: the
INPE-EM framework applied to the Brazilian Amazon. Glob. Chang. Biol. 18,
3346–3366.
ANEEL — Agência Nacional de Energia Elétrica, 2014. Banco de Informação de Geração.
(Available at: http://www.aneel.gov.br/aplicacoes/capacidadebrasil/capacidadebrasil.
cfm).
Antunes, E., 2009. Perspectivas da geração termelétrica a carvão no Brasil no horizonte
2010–2030. (Master Thesis). COPPE/UFRJ, Planejamento Energético, Rio de Janeiro,
Brazil.
Bermann, C., 2012. O projeto da Usina Hidrelétrica Belo Monte: a autocracia energética
como paradigma. Novos Cadernos. 15. NAEA, pp. 5–23. ISSN: 1516-6481.
Boden, T., Marland, Greeg, Andres, Bob, 2013. Global, Regional, and National Fossil-Fuel
CO2 Emissions. Carbon Dioxide Information Analysis Center, Oak Ridge National Lab-
oratory, US Department of Energy, Oak Ridge, TN (Available at: http://cdiac.ornl.gov/
trends/emis/overview_2013.html).
Borba, B.S.M.C., Szklo, A.S., Schaeffer, R., 2012. Plug-in hybrid electric vehicles as a way to
maximize the integration of variable renewable energy in power systems: the case of
wind generation in northeastern Brazil. Energy 37, 469–481.
BP, 2014. BP Statistical Review of World Energy (Available at: bp.com/statisticalreview).
Branco, D.A.C., Moura, Maria Cecilia P., Szklo, Alexandre, Schaeffer, Roberto, 2013. Emis-
sions reduction potential from CO2 capture: a life-cycle assessment of a Brazilian
coal-fired power plant. Energ. Policy 61, 1221–1235 (October 2013).
Brasil, 2009. Decree No 12.187, December 29th 2009. (Available at: http://www.planalto.
gov.br/ccivil_03/_ato2007-2010/2009/lei/l12187.htm).
Burgi, A., 2013. Avaliação do Potencial Técnico de Geração Elétrica Termossolar no Brasil a
partir de Modelagem em SIG e Simulação de Plantas Virtuais. (Master Thesis). COPPE/
UFRJ, Planejamento Energético, Rio de Janeiro, Brazil.
Cabrera-Palmer, B., Rothwell, Geoffrey, 2008. Why is Brazil enriching uranium? Energ.
Policy 36, 2570–2577.
Calvin, K.V., et al., 2011. GCAM Wiki Documentation. (https://wiki.umd.edu/gcam/).
Calvin, K.V., et al., 2016. Agriculture, forestry, and other land-use emissions in Latin America.
Energy Econ. 56, 615–624 (in this issue).
Camargo, T.R.M., Merschmann, P.R.C., Arroyo, E.V., Szklo, A., 2014. Major challenges for
developing unconventional gas in Brazil — will water resources impede the develop-
ment of the Country's industry? Resour. Policy 41, 60–71.
Carvalho, J.F., Sauer, I.L., 2009. Does Brazil need new nuclear power plants? Energ. Policy
37, 1580–1584.
CEPEL — Centro de Pesquisas de Energia Elétrica, 2001. Atlas do Potencial Eólico Brasileiro.
CEPEL, Rio de Janeiro, Brazil.
Clarke, L., Krey, Volker, Weyant, John, Chaturvedi, Vaibhav, 2012. Regional energy system
variation in global models: results from the Asian Modeling Exercise scenarios. Ener-
gy Econ. 34 (Supplement 3), S293–S305 (December 2012).
Clarke, L., McFarland, J., Octaviano, C., van Ruijven, B., Beach, R., Daenzer, K., Hernandez, S.,
Lucena, A.F.P., Kitous, A., Labriet, M., Rodriguez, A.M.L., Mundra, A., van der Zwaan, B.,
2016. Long-term abatement potential and current policy trajectories in Latin Ameri-
can countries. Energy Econ. 56, 513–525 (in this issue).
Costa, I.V.L., 2014. Proposta de Estrutura Regulatória para Sequestro Geológico de CO2 no
Brasil e uma Aplicação para o estado do Rio de Janeiro. (DSc. Dissertation). COPPE/
UFRJ, Planejamento Energético, Rio de Janeiro, Brazil.
Costa, A.D., Souza-Santos, E.R., 2009. As jazidas petrolíferas do pré-sal: marco regulatório,
exploração e papel da Petrobras. Trabalho periódico 0096. Universidade Federal do
Paraná, Departamento de Economia.
Denmark, 2009. Overview of the Brazilian oil and gas industry. Report Prepared for
Offshore Center Denmark (Available at: http://www.offshorecenter.dk/filer/files/
Project/OCD%20repor/OCD%20report%20Brazil.pdf).
DNPM, 2013. Sigmine - Sistema de Informações Geográficas da Mineração. Available at:.
http://sigmine.dnpm.gov.br/webmap/.
EIA — U.S. Energy Information Administration, 2011. World Shale Gas Resources: An Ini-
tial Assessment of 14 Regions Outside the United States. (Available at: http://www.
eia.gov/analysis/studies/worldshalegas/).
Eletronuclear, 2011. Expansão Nuclear no Nordeste. (Available at: http://cnn.
eletronuclear.gov.br).
EMBRAPA — Empresa Brasileira de Pesquisa Agropecuária, 2009. Zoneamento
Agroecológico da Cana de Açúcar, Empresa Brasileira de Pesquisa Agropecuária.
p. 59 (Rio de Janeiro, RJ).
EPE — Empresa de Pesquisa Energética, 2007. Plano Nacional de Energia 2035. (Rio de
Janeiro, 2007. Available at: http://www.epe.gov.br).
EPE — Empresa de Pesquisa Energética, 2008. Plano Nacional de Energia 2030 – Geração
Hidrelétrica. Empresa de Pesquisa Energética. MME: EPE, Brasília (Available at:
http://epe.gov.br/PNE/Forms/Empreendimento.aspx).
EPE — Empresa de Pesquisa Energética, 2013. Plano Decenal de Expansão de Energia
2022. (Rio de Janeiro, 2013. Available at: http://www.epe.gov.br).
EPE — Empresa de Pesquisa Energética, 2014. National Energy Balance, 2014. (Available
at: https://ben.epe.gov.br/).
Farias, L.T., 2014. Avaliação da Produção Integrada de Hidrogênio e Metanol para Redução
de Emissões de Carbono no Refino de Petróleo. (Master Thesis). COPPE/UFRJ,
Planejamento Energético, Rio de Janeiro, Brazil.
Fishman, A.D., 2010. Petroleum in Brazil: Petrobras, petro-sal, legislative changes and the
role of foreign investment. Trabalho periódico — the George Washington University
(Available at: bhttp://www.gwu.edu/~clai/working_papers/Fishman_Andrew_12-
10.pdfN).
574 A.F.P. Lucena et al. / Energy Economics 56 (2016) 564–574
Fugimoto, S., 2010. Estrutura de tarifas de energia elétrica: análise crítica e proposições
metodológicas. (DSc. Dissertation). USP, São Paulo, Brazil.
Gavronski, J., 2007. Carvão mineral e as energia renováveis no Brasuil. (DSc. Dissertation).
UFRGS, Porto Alegre, Brazil.
GEA, 2012. Global Energy Assessment — Toward a Sustainable Future. Cambridge Univer-
sity Press, Cambridge, UK and New York, NY, USA and the International Institute for
Applied Systems Analysis, Laxenburg, Austria.
Griffin, B., et al., 2014. White Knights: will wind and solar come to the rescue of a looming
capacity gap from nuclear phase-out or slow CCS start-up? Clim. Chang. 123, 623–635.
GWEC — Global Wind Energy Council, 2011. Global Wind Report: Annual Market Update
2011. Global Wind Energy Council, Brussels.
Hoffman, B.S., Szklo, A., 2011. Integrated gasification combined cycle and carbon capture:
a risky option to mitigate CO2 emissions of coal-fired power plants. Appl. Energy 88
(11), 3917–3929.
Hoffman, et al., 2012. Simulação dos impactos da implantação de sistemas de captura de
CO2 sobre os custos, a geração de resíduos e o consumo de H2O de termelétricas a
carvão. Congresso Brasileiro de Carvão Mineral.
Hoffmann, B.S., 2010. O ciclo combinado com gaseificação integrada e a captura de CO2:
uma solução para mitigar as emissões de CO2 em Termelétricas a carvão em larga
escala no curto prazo. (Master Thesis). COPPE/UFRJ, Planejamento Energético, Rio
de Janeiro, Brazil.
Hoffmann, B.S., 2013. O potencial termelétrico a carvão no Rio Grande do Sul diante
restrições de disponibilidade de água e objetivos de redução de emissões de CO2,
aplicando a queima em leito fluidizado/Bettina Susanne Hoffmann. (DSc. Disserta-
tion). COPPE/UFRJ, Planejamento Energético, Rio de Janeiro, Brazil.
Hoffmann, B.S., Szklo, A., 2014. Limits to co-combustion of coal and eucalyptus due to
water availability in the state of Rio Grande do Sul, Brazil. Energy Conversion and
Management (In Press, Corrected Proof).
IAEA (International Atomic Energy Agency), 2006. Brazil: a country profile on Sustainable
Energy Development, Vienna.
IEA — International Energy Agency, 2013. World Energy Outlook. IEA, Paris.
Imperio, M., Rochedo, P., Costa, I., Szklo, A., 2014. CO2 capture for associated natural gas in
Brazil's pre-salt basins. Technical Report to Projeto Opções de Mitigação. Programa de
Planejamento Energético/COPPE/UFRJ.
IPCC — Intergovernmental Panel on Climate Change, 2014. Climate Change 2014: Mitiga-
tion of Climate Change. Contribution of Working Group III to the Fifth Assessment
Report of the Intergovernmental Panel on Climate Change. In: Edenhofer, O., Pichs-
Madruga, R., Sokona, Y., Farahani, E., Kadner, S., Seyboth, K., Adler, A., Baum, I.,
Brunner, S., Eickemeier, P., Kriemann, B., Savolainen, J., Schlömer, S., von Stechow, C.,
Zwickel, T., Minx, J.C. (Eds.), Cambridge University Press, Cambridge, United
Kingdom and New York, NY, USA.
Kitous, et al., 2010. Transformation patterns of the worldwide energy system — scenarios
for the century with the POLES model. Energy J. vol. 31, 49–82 (Special Issue).
Kober, T., van der Zwaan, B.C.C., Rösler, H., 2014. Emission certificate trade and costs
under regional burden-sharing regimes for a 2 °C climate change control target.
Clim. Chang. Econ. 5 (1), 1–32 (1340013).
Lage, E.S., Processi, L.C., Souza, L.D.W., Dores, P.B., Galoppi, P.P.S., 2013. Gás não
convencional: experiência Americana e perspectivas para o mercado brasileiro.
Petróleo e Gás 37. BNDES Setorial, pp. 33–88.
Larson, E.D., 2006. A review of lifecycle analysis studies on liquid biofuel systems for the
transport sector. Energy Sustain. Dev. X (2), 109–126.
Leal, M.R.L.V., Nogueira, L.A.H., Cortez, L.A.B., 2013. Land demand for ethanol production.
Appl. Energy 102, 266–271.
Lucena, A.F.P., Szklo, A.S., Schaeffer, R., Souza, R.R., Borba, B.S.M.C., Costa, I.V.L., Pereira
Júnior, A.O., Cunha, S.H.F., 2009. The vulnerability of renewable energy to climate
change in Brazil. Energy Policy 37, 879–889. http://dx.doi.org/10.1016/j.enpol.2008.
10.029.
Lucena, A.F.P., Szklo, A.S., Schaeffer, R., 2010. Least-cost adaptation options for global cli-
mate change impacts on the Brazilian electric power system. Glob. Environ. Chang.
20, 342–350.
Lucon, O., Romeiro, V., Pacca, S., 2013. Reflections on the international climate change ne-
gotiations: a synthesis of a working group on carbon emission policy and regulation
in Brazil. Energ. Policy 59, 938–941.
Malagueta, D., Szklo, Alexandre, Soria, Rafael, Dutra, Ricardo, Schaeffer, Roberto, Bruno
Soares Moreira Cesar Borba, 2014. Potential and impacts of Concentrated Solar
Power (CSP) integration in the Brazilian electric power system. Renew. Energy vol.
68, 223–235. http://dx.doi.org/10.1016/j.renene.2014.01.050 (August 2014).
MAPA — Ministério da Agricultura, Pecuária e Abastecimento, 2014. Lifestock and Agricul-
tural Plan 2013/2014 [Plano agrícola e pecuário 2013/2014] [in Portuguese]. Ministry
of Agriculture, Government of Brazil, Brasilia.
Mathias, M.C., Szklo, A., 2007. Lessons learned from Brazilian natural gas industry reform.
Energ. Policy 35, 6478–6490.
Matthews, H.D., Graham, Tanya L., Keverian, Serge, Lamontagne, Cassandra, Seto, Donny,
Smith, Trevor J., 2014. National contributions to observed global warming. Environ.
Res. Lett. 9.
Maugeri, L., 2012. Oil: the next revolution. Discussion Paper #2012-10. Belfer Center for
Science and International Affairs, Harvard Kennedy School.
Merschmann, P.R.C., Vasquez, Eveline, Szklo, Alexandre S., Schaeffer, Roberto, 2013.
Modeling water use demands for thermoelectric power plants with CCS in selected
Brazilian water basins. Int. J. Greenhouse Gas Control vol. 13, 87–101 (March 2013).
Miranda, R.F.C., Szklo, A.S., Schaeffer, R., 2014. Technical–economic potential of PV sys-
tems on Brazilian rooftops. Renew. Energy http://dx.doi.org/10.1016/j.renene.2014.
10.037 (submitted for publication).
Nassar, A.M., Harfuch, L., Bachion, L.C., Moreira, M.M., 2011. Biofuels and land-use chang-
es: searching for the top model interface. Focus 1, 224–232.
Nepstad, D., McGrath, David, Stickler, Claudia, Alencar, Ane, Azevedo, Andrea, Swette,
Briana, Bezerra, Tathiana, DiGiano, Maria, Shimada, João, Seroa da Motta, Ronaldo,
Armijo, Eric, Castello, Leandro, 2014. Slowing Amazon deforestation through public
policy and interventions in beef and soy supply chains. Science vol. 344 (6188),
1118–1123 (6 June 2014).
Nogueira, L.P.P., Lucena, A.F.P., Rathmann, R., Rochedo, P.R.R., Szklo, A., Schaeffer, R., 2014.
Will thermal power plants with CCS play a role in Brazil's future electric power gen-
eration? Int. J. Greenhouse Gas Control 24, 115–123 (May 2014).
OCD — Offshore Center Denmark, 2009. Overview of the Brazilian oil and gas industry.
Report — Offshore Brazil (Available at: http://www.offshorecenter.dk/filer/files/
Project/OCD%20repor/OCD%20report%20Brazil.pdf).
Ortiz, 2011. Avaliação Técnico-Econômica de Sistemas IGCC Utilizando Coque de Petróleo
e Carvão Mineral como Combustível. (Master Thesis). UNIFEI, Itajubá, Brazil.
Paltsev, S., Reilly, J., Jacoby, H., Eckaus, R., Mcfarland, J., Sarofim, M., Asadoorian, M.,
Babiker, M., 2005. The MIT Emissions Prediction and Policy Analysis (EPPA) model:
version 4. MIT Joint Program on the Science and Policy of Global Change, Report
125. MIT, Cambridge, MA (Available at: http://globalchange.mit.edu/research/
publications/697).
Paltsev, S., Monier, E., Scott, J., Sokolov, A., Reilly, J., 2013. Integrated economic and climate
projections for impact assessment. Clim. Chang. http://dx.doi.org/10.1007/s10584-
013-0892-3.
Peters, G.P., Marland, Gregg, Le Quéré, Corinne, Boden, Thomas, Canadell, Josep G.,
Raupach, Michael R., 2012. Rapid growth in CO2 emissions after the 2008–2009 global
financial crisis. Nat. Clim. Chang. 2, 2–4.
Petrobras, 2011. Relatório de Sustentabilidade. (Available at: http://www.petrobras.com.
br/pt/sociedade-e-meio-ambiente/relatorio-de-sustentabilidade/).
Pimentel, T., 2012. O Enfrentamento Político dos Conflictos Socioambientais Decorrentes da
Implantação de Usinas Hidrelétricas. (MSc. Thesis), Universidade Católica de Brasilia
(Available at:http://www.aneel.gov.br/biblioteca/trabalhos/trabalhos/Dissertacao_
Tamara_Pimentel.pdf).
Portugal-Pereira, J., Soria, R., Rathmann, R., Schaeffer, R., Szklo, A., 2014. Techno-economic
and environmental assessment of agricultural and agro-industrial waste-to-energy
potential in Brazil. Paper Presented in 5th International Conference on Engineering
for Waste and Biomass Valorisation — August 25–28 (Rio de Janeiro, Brazil).
Rochedo, P.R.R., Szklo, A., 2013a. Designing learning curves for carbon capture based on
chemical absorption according to the minimum work of separation. Appl. Energy
108 (8), 383–391.
Rochedo, P.R.R., Szklo, A.S., 2013b. Economic analysis under uncertainty of coal fired
capture-ready power plants. Int. J. Greenhouse Gas Control vol. 12, 44–55 (January
2013).
Saraiva, T., Szklo, A., Lucena, A., Chavez-Rodriguez, M.F., 2014. Forecasting Brazil's crude
oil production using a multi-Hubbert model variant. Fuel 115, 24–31.
Simões, R., 2010. Seminário no Brazil Wind Power 2010, Associação Brasileira de Energia
Eólica (ABEEólica) (Rio de Janeiro, RJ, Brazil).
Singer, C., 2013. U.S. spent nuclear fuel management: political, fiscal, and technical feasi-
bility. Energ. Policy 61, 1521–1528 (October 2013).
SolarGIS, 2013. SolarGIS. iMaps. Solar Data 2013 (Available at: http://solargis.info/).
UNEP — United Nations Environmental Programme, 2013. The Emissions Gap Report
2013. United Nations Environmental Programme (UNEP), Nairobi.
USGS — U.S. Geological Survey, 2000. US World Geological Survey Petroleum Assessment
(Washington, DC).
van der Zwaan, B.C.C., 2013. The role of nuclear power in mitigating emissions from elec-
tricity generation. Energy Strateg. Rev. 1, 296–301.
van der Zwaan, B.C.C., Rösler, H., Kober, T., Aboumahboub, T., Calvin, K.V., Gernaat, D.E.H.J.,
Marangoni, G., McCollum, D.L., 2014. A cross-model comparison of global long-term
technology diffusion under a 2 °C climate change control target. Clim. Chang. Econ.
4 (4), 1–24 (2013, 1340013).
van der Zwaan, L., Clarke, K., Calvin (Guest Editors), 2016a. Climate Policy in Latin Amer-
ica: implications for energy and land use. Overview of a Special Issue on the findings
of the CLIMACAP-LAMP project, Energy Econ. 56, 495–498 (in this issue).
van der Zwaan, B., et al., 2016b. Energy technology roll-out for climate change mitigation:
a multi-model study for Latin America. Energy Econ. 56, 526–542 (in this issue).
van Ruijven, B., et al., 2016. A cross-model investigation of energy and emission baseline
scenarios for Latin America. Energy Econ. 56, 499–512 (in this issue).
Wing, I.S., Daenzer, K., Fisher-Vanden, K., Calvin, K., 2011. Phoenix Model Documentation.
(Available at: http://www.globalchange.umd.edu/models/phoenix/).
Wiser, R., Yang, Z., Hand, M., Hohmeyer, O., Infield, D., Jensen, P.H., Nikolaev, V., O'Malley,
M., Sinden, G., Zervos, A., 2011. Wind energy. In: Edenhofer, O., Pichs-Madruga, R.,
Sokona, Y., Seyboth, K., Matschoss, P., Kadner, S., Zwickel, T., Eickemeier, P., Hansen,
G., Schlömer, S., von Stechow, C. (Eds.), IPCC Special Report on Renewable Energy
Sources and Climate Change Mitigation. Cambridge University Press, Cambridge,
United Kingdom and New York, NY, USA.