Renewable and Sustainable Energy Reviews 72 (2017) 1124–1137

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

journal homepage: www.elsevier.com/locate/rser

A foundation for the strategic long-term planning of the renewable energy MARK
sector in Brazil: Hydroelectricity and wind energy in the face of climate
change scenarios
Vera Ruffato-Ferreiraa, Renata da Costa Barretob, Antonio Oscar Júniorc,
⁎
Wanderson Luiz Silvad, Daniel de Berrêdo Vianab, , José Antonio Sena do Nascimentoe,
Marcos Aurélio Vasconcelos de Freitasa
a
Federal University of Rio de Janeiro (UFRJ), Energy Planning Program (EPP/COPPE), Pedro Calmon Avenue, s/n° – Centro de Tecnologia, block C, Room
216, University City, Ilha do Fundão, Rio de Janeiro, RJ, Brazil
b
International Virtual Institute of Global Change (IVIG/COPPE/UFRJ), Avenida Pedro Calmon, s/n°, Bloco P – P2, Cidade Universitária - Ilha do Fundão,
CEP: 21941-596 Rio de Janeiro, RJ, Brazil
c
State University of Rio de Janeiro, Institute of Geography, Avenida São Francisco Xavier, 524, 4th floor, block F, CEP: 20550-900 Rio de Janeiro, RJ, Brazil
d
Centre of Mathematical and Nature Sciences (CCMN), Institute of Geosciences (IGEO), Department of Meteorology, Avenida Athos da Silveira Ramos, 274,
Bloco A, Cidade Universitária - Ilha do Fundão, CEP: 21941-916 Rio de Janeiro, RJ, Brazil
e
Brazilian Institute of Geography and Statistics (IBGE), Coordination of Natural Resources and Environmental Studies (CREN/DGC), República do Chile
Avenue, 500, 15th floor, Downtown, Rio de Janeiro, RJ, Brazil

A R T I C L E I N F O A BS T RAC T

Keywords: The current article assesses the vulnerability of renewable energies in Brazil in the face of climate change, with
Renewable energy hydroelectricity and wind power taken as case studies. Future climate data produced by the National Institute
Hydroelectricity for Space Research were employed for three time slices (2010–2040, 2041–2070 and 2071–2100) taken from
Wind energy the Eta 20 km regional model using the global model HadGem2-ES of the IPCC AR5, for scenarios RCP 4.5 and
Brazil
8.5. Based on these data and using indicators, it was possible to generate predictions that are useful for the
Climate modeling
evaluation of the impacts of future climate scenarios on the current energy infrastructure, enabling the
proposition of recommendations for mitigation and adaptation in the energy sector, a strategic subject for
Brazilian governmental programs.

1. Introduction hydroelectricity. Approximately 40% of the internal energy supply is

With the current energy scenario in Brazil and the water crisis in its
South and Southeast regions [1], discussing the planning of the energy
sector has never been more decisive for the energy security of the
country, as it ensures the maintenance of sustainable development by
providing a foundation for the planning of governmental programs.
The priority water use is given to human and animal consumption,
defined by the National Water Resources Policy. The lack of priority on
water use by other sectors of the economy, along with increased
demand, have increased the conflict of interest, especially in cases of
water deficits [2].
According to the Energy Research Company [3], renewable energy
have a big stake in the Brazilian energy matrix, with a predominance of
from renewable sources. By 2030, 45% of all energy consumption is
expected to come from renewable sources.
Under the Brazilian strategic planning, energy takes preponderant
role in the interconnection with all sectors of the economy and social
welfare. In this sense, the National Energy Plan (NEP), the Decennial
Plan for Energy Expansion (DPE), both subsidized by the National
Energy Balance (BEN) data, are instruments that embody the Brazilian
energy planning. The NEP aims to provide inputs for the energy
planning of the country in the long term. The last NEP fully published
makes projections of energy demand and economic scenarios to 2030
[2].
The DPE is an annual basis, prepared by the Energy Research
Company (EPE), under the supervision of the Ministry of Mines and

⁎
Corresponding author.
E-mail addresses: veraruffato@ivig.coppe.ufrj.br (V. Ruffato-Ferreira), rbarreto@ivig.coppe.ufrj.br (R. da Costa Barreto), antonio.oscar@ivig.coppe.ufrj.br (A. Oscar Júnior),
wanderson@ufrj.br (W.L. Silva), danberredo@ivig.coppe.ufrj.br (D. de Berrêdo Viana), joseantoniosena@gmail.com (J.A.S. do Nascimento),
mfreitas@ivig.coppe.ufrj.br (M.A.V. de Freitas).

http://dx.doi.org/10.1016/j.rser.2016.10.020
Received 8 October 2015; Received in revised form 19 July 2016; Accepted 16 October 2016
Available online 31 October 2016
1364-0321/ © 2016 Elsevier Ltd. All rights reserved.
V. Ruffato-Ferreira et al.
Energy (MME) and presents the expected demand and supply of energy
for the next ten years, including the disclosure annual energy matrixes
[4].
Renewable energy strongly depends on climate conditions.
However, the long-term energy planning in Brazil does not consider
the vulnerability of renewable energy sources to global climate change
[5]. The monitoring of climate variations through vulnerability analysis
is important as a method to create mitigation and adaptation strategies
of the Brazilian energy sector in face of these changes, with no or little
impact on the socioeconomic development of the country.
The diversification of the matrix is fundamental to ensure energy
security in the context of the current energy crisis and to avoid similar
situations in the future. In a study aimed at long-term planning, it is
necessary to consider the effects of the climate changes that will
significantly affect the geographical distribution and the potential for
energy generation of the country.
Several studies show the Brazilian observed trends over the past
decades in climate variables, especially precipitation and air tempera-
ture. At large, these trends point to a general warming in most parts of
the country, as well as an increase in the frequency and intensity of
extreme precipitation events [6–8]. Still in relation to climate change
research, several studies provide assessment over future projections of
these variables over the twenty-first century, mainly pointing out rising
temperatures, with increased events of severe rainfall and droughts in
some regions [9–11].
The seven largest producers of hydroelectricity in the world are
China, Canada, Brazil, the United States, Russia, India and Norway.
Brazil is the second largest producer, accounting for 12% of the total,
and China is the first with 19.6% of worldwide production [12].
Hydroelectric production corresponds to 11.5% of the internal energy
supply in Brazil, with 65.2% of the supply of national electricity [3].
Much of the hydroelectric potential of the South, Southeast and
Northeast has already been explored, concentrating 80% of installed
capacity. Most of the potential to be exploited is concentrated in the
Mid-West and North (Amazon). The expansion plans of the sector have
turned to exploiting the potential of river basins of the Amazon region,
including large projects, such as plants of Belo Monte, Santo Antônio
and Jirau, in addition to the efforts in linking this region with the
distribution of the National Interconnected System (NIS) [2,13].
Conditions of hydroclimatic variability, especially those extreme
and/or unexpected, such as droughts and floods, can affect or even
interrupt power generation in a hydroelectric power plant. As it is
nearly impossible to predict these situations and therefore to adapt and
respond to them, social, economic and ecological impacts can occur. A
significant part of hydrological variability is stimulated by climate
dynamics [14,15].
The existing regulation plans are almost entirely based on historical
hydrological records, and do not consider new models neither advances
in the understanding of the climate system in many time scales,
including the future; a flood, for instance, can occur within hours
and a drought in the period of months of years [14]. In a scenario of
global climate change, which will affect the average behavior of
watershed discharges, we can predict medium and long-term losses
in the operation of the hydroelectric system [15]. These considerations
should be taken into account in the decision-making process, for they
may offer both risks and new opportunities for hydroelectric systems
[14].
Milly et al. [16] analyzed the rivers streamflow of nine projection
models from IPCC AR4. For the mid-twenty-first century (2041–2060),
the models projected reductions in São Francisco, Parnaíba, Tocantins,
Xingu rivers streamflow and others in the eastern Amazon. Lucena
et al. [5] examined the interaction between future climate change and
renewable energy in Brazil, including hydroelectricity. The authors
used the HadCM3 global climate model precipitation data in relation to
scenarios A2 and B2 of the IPCC AR4. The results indicate that the
average annual streamflow (that is, the average annual amount of water
1125
Renewable and Sustainable Energy Reviews 72 (2017) 1124–1137
that flows into the hydroelectric plants) may suffer an average decline
of 8.6% in the A2 scenario and 10.8% in the B2 scenario by the end of
XXI century. Bravo et al. [17] conducted hydrologic simulations using
data from climate models between 1961 and 2099. The results show
that in the South region and in the southern part of the Paraná River
basin, there is a larger number of climate models suggesting the rise of
streamflow rates [17].
Hydroelectric plants with large reservoirs are the most adopted in
Brazil. Reservoirs reduce the vulnerability of the energy matrix as they
have enough storage capacity to keep generating in cases of minor
changes in rainfall patterns. However, with increasing consideration of
environmental issues, the new plants that are being installed are based
on run-of-the-river hydropower plant, with smaller reservoirs [2,5].
Thus, hydropower production becomes more vulnerable to climatic
variations and the increase of extreme events resulting from global
climate change.
As for wind energy, the five largest wind power generators in the
world are China, the United States, Germany, Spain and India.
According to BIG ANEEL database, Brazil has 336 wind farms with
8.2 GW operating and 3.5 GW under construction as on February
2016. This amount corresponds to approximately 6.07% of the
electricity production capacity and 35.31% of the production capacity
of projects under construction [18]. The biggest wind potential in
Brazil is in the Northeast, also where most of the installed plants are
located. The government has strongly encouraged the sector expansion,
however, the price of the turbines also require a very high investment,
limiting some projects [2].
Detailed knowledge of wind behavior is required for the analysis of
wind potential, owing to the correlation between both. Speed intensity
and wind direction are the main data for the determination of the wind
potential of a region, and they are related to relief, soil roughness and
other obstacles distributed throughout the assessed area. For wind
power to be considered technically exploitable, its density needs to be
higher than or equal to 500 W/m2 at a 50-meter height, which requires
a minimum wind speed of 7–8 m s−1 [19].
Many studies and surveys on the topic have stimulated the
commercial exploitation of the Brazilian wind potential. The first
studies were carried out in the States of Ceará and Pernambuco, and
the Brazilian Center of Wind Energy of the Federal University of
Pernambuco published the first version of the Wind Atlas of the
Northeast Region, with support from ANEEL and the Ministry of
Science, Technology and Innovation. Moreover, the Overview of the
Wind Potential in Brazil is a result of the continuity of this work.
The highest wind potentials were identified along the coast of the
Northeast region and in the South and Southeast regions [20,21]. The
annual energy potential is about 144.29 TW h/year for the Northeast
region, 54.93 TW h/year for the Southeast and 41.11 TW h/year for
the South [22]. The study of the variation of the country's potential for
wind power production due to climate change is important due to the
existing potential, to recent technological innovations that increased its
competitiveness within the energy market and to the need for a more
diverse energy matrix to reduce the sector's vulnerability [23,24].
Lucena et al. [24] examined some possible impacts of climate
change on the Brazilian wind energy potential. According to the
projections, expectations are for better conditions for wind energy
generation on the coast of the Northeast region. Pereira et al. [23]
projected Brazilian wind speed in the Eta-HadCM3 model. The A1B
scenario simulations indicate an increase trend from 15% to 30% in
wind energy density for most of the Northeast and above 100% in the
northern sector of the region.
During the 21th Conference of the Parties held by the UN in
November 2015 (COP 21 - Paris), Brazil presented its Intended
Nationally Determined Contributions (INDCs), in which it undertook
to make efforts to a “transition to renewable energy systems and
decarbonizing the economy” through a 43% reduction in emissions of
greenhouse gases (GHG) emissions by 2030 compared to 2005 levels
V. Ruffato-Ferreira et al.
and achieve 45% renewable energy (including hydropower) [25].
Considering these perspectives, the current work presents an
analysis of this vulnerability and an interpretation of the positive and
negative future climate changes on the sector of renewable energies
(hydropower plants and wind power), with the purpose of providing
information for researches and public policies aimed at adaptations
and implementations in the current energy matrix of Brazil.
2. Methodology
2.1. Climate model data
As explained by Moss et al. [26], the climate research communities
have cooperated to create and use specific emission scenarios as a
plausible pathway towards reaching each target radiative forcing1
trajectory. They are based on the identification of important character-
istics (including data on land use and land cover) for climate modeling
until year 2100. These radiative forcing trajectories are not associated
with unique socioeconomic or emissions scenarios, and instead can
result from different combinations of economic, technological, demo-
graphic, policy and institutional futures.
In order to calculate all indicators and indexes related to the
vulnerability of renewable energies in the face of global climate change
that were proposed in this study, the averages of thirty years were used
as reference for the treatment of the data produced by the National
Institute for Space Research – INPE, in the following time slices:
present (1961–90) and future (2010–40, 2041–70 and 2071–2100),
taken from the Eta 20 km regional model using the global model
HadGem2-ES of the IPCC AR5, for the RCPs 4.5 and 8.5 [27].
The RCP 4.5 scenario considers as a premise the medium-low
radiative stabilization force, arising from climate policies, such as the
introduction of global prices of GHG emissions, invoked to achieve the
goal of limiting the radiative forcing to 4.5 W/m2, without exceeding
this value in the twenty-first century, related to the old B1 scenario
[28–30]. The RCP 8.5 does not include any specific purpose of
mitigation. GHG emissions and concentrations of this scenario sig-
nificantly increase over time, leading to a radiative forcing of 8.5 W/m2
at the end of the century. Assumptions relating to trends or on
technological, demographic and economic changes in society are based
on the revised and expanded SRES A2 scenario, but end up with a
higher forcing in 2100 close to the A1FI scenario [28,30,31].
The Eta model corresponds to an atmospheric model of limited area
developed at the University of Belgrade (Serbia), and implemented by
the National Center for Environmental Predictions (National Centers
for Environmental Prediction – NCEP) in the United States [32]. The
model was chosen because it uses the Eta (η) coordinate system, which
reduces calculation errors when there is insertion of vertical coordi-
nates, significant in the case of the South American continent due to
the presence of the Andes [33]. This Eta model was nested in the global
model HadGEM2-ES, now denominated Eta-HadGEM2-ES, using
their boundary conditions for the purpose of generating data with
better resolution at regional level.
The Eta-HadGEM2-ES application to Brazil was conducted by the
Weather Forecasting and Climate Studies Center of the National
Institute for Space Research (CPTEC/INPE). This was the only model
available to the scenarios of interest. Although some specific errors,
simulations of this model can satisfactorily represent the climatological
aspects in South America [35,36]. The Eta-HadGEM2-ES model has a
more refined resolution grid (20 km) compared to previous regionali-
zation of the same model, highlighting the climate characteristics in
Brazil that would not be possible with a rougher global model.
1
Radiative forcing is the change in the net, downward minus upward, irradiance
(expressed in Watts per square meter, W/m2) at the tropopause due to a change in an
external driver of climate change, such as, for example, a change in the concentration of
carbon dioxide or the output of the Sun, usually relative to the year 1750 [34].
1126
Renewable and Sustainable Energy Reviews 72 (2017) 1124–1137
Moreover, these new simulations incorporate the most recent climate
change scenarios of the IPCC AR5, the RCPs [37].
2.2. Hydroelectricity Indicator
Hydroelectricity production depends on the water cycle, in which
part of the rainwater that reaches the soil runs superficially and is
drained to water bodies of the drainage basin, influencing river
discharge, which is used to generate electricity through hydropower
plants, among other applications. To evaluate the vulnerability of the
Brazilian hydroelectricity sector in the face of global climate change, we
employ two indexes: the Water Balance Index, to predict the variation
of watershed discharges in Brazil in relation to precipitation (PREC),
and the Evapotranspiration Index, which estimates evapotranspiration
rates from latent heat flux.
The Evapotranspiration Index is based on the precept that changing
the physical state of water from liquid to gas demands energy
consumption of 590 cal/g, which allows us to estimate the transforma-
tion of latent heat flux into evapotranspiration [38]. For that purpose,
the following conversion factors are employed: 1 W/m2=0.0864 MJ/
m2 day and 1 mm/day=2.45 MJ/m2 day [39].
The calculation of the Water Balance index will be used to correlate
precipitation with evapotranspiration, allowing us to assess the in-
creasing or decreasing trend in water availability in Brazilian drainage
basins, and consequently their hydroelectric production.
The equation employed to calculate the Water Balance index (WB)
is:
WB = PREC − EVAP
where:
WB=Water Balance (mm/day).
PREC=average daily precipitation in a period of 30 years (mm/day).
EVAP=evapotranspiration index (mm/day).
The WB indexes are presented in maps created by interpolation,
using a geoprocessing software. The maps indicate a positive or
negative variation of the index (making the region more humid or
drier) per watershed. For that purpose, we employed the watershed
division of the National Department of Waters and Electric Power –
DNAEE (Fig. 1). The ratio of the water balance is directly proportional
to the river discharge, affecting the hydroelectric potential of the region
as mentioned above.
2.3. Wind energy indicator
Wind power generation depends on turbines that capture part of
the kinetic energy of the wind, which is then transformed into electric
energy as it passes through the swept area of the rotor [41]. As a result,
the production of wind power is related to wind behavior, which is
determined mainly by its speed intensity and direction.
For wind power to be considered technically exploitable, its density
needs to be equal to or higher than 500 W/m2 at the height of the
turbines, requiring a minimum wind speed of 7 or 8 m s−1 [19]. For
Schaeffer et al. [15], on the other hand, the minimum speed for the
exploitation of wind power is 6 m s−1, and values above 8.5 m s−1 are
considered high.
In this context, the average total wind speed index and the wind
intensity trend indicator per season will be used to predict wind
variations in the entire territory of Brazil, with the purpose of assessing
the vulnerability of the country's wind sector in the face of global
climate change.
The total wind speed is a vector composition given by its zonal and
meridional components. As new wind turbines tend to be installed at a
height between 80 and 100 m, the wind speed at an altitude of 100 m
will be used. The index will be calculated using the following equation:
V. Ruffato-Ferreira et al. Renewable and Sustainable Energy Reviews 72 (2017) 1124–1137

Fig. 1. Watershed Division of the DNAEE [40].

Fig. 2. Distribution of the water balance in Brazil for the historical scenario and future scenarios, using the Eta-HadGEM2 model and scenarios RCP 4.5 and 8.5.

⇀
⎯ with interpolation, using a geoprocessing software. To provide the
V = U100 2 + V 100 2
analysis of the variations of areas with different wind intensities, wind
where: speeds were divided into classes of higher or smaller wind production:
< 3 m s−1, 3–4 m s−1, 4–5 m s−1, 5–6 m s−1, 6–7 m s−1, 7–8 m s−1, >
U100=zonal wind at an altitude of 100 m. 8 m s−1.
V100=meridional wind at an altitude of 100 m.

It is important to consider that wind progression and speed are
influenced by irregularities of the terrain, characteristics of the relief
and patterns of vegetation [15]. In spite of the importance of the
impact of vegetation on the climate projections of wind speeds, it was
not possible to consider any projection of vegetation based on the
climate projections of the model. Consequently, the variable will not be
included in this study, which will present a simplified evaluation of
total wind speed. The influence of vegetation on wind speed should be
considered in future analyses.
The Total Wind Speed indexes will be presented in maps created
2.4. Statistical tests
The Mann Kendall non-parametric statistical test [42] was used to
evaluate the significance of the climate trends. The test has been
considered the most appropriate method to analyze the statistical
significance of possible climate changes in climatological time series
[43].
In this article, we also calculate Sen's Bend, which is a non-
parametric method (assuming a linear trend) used to estimate the
magnitude of trends [44]. As Sen's Bend is insensitive to outliers and

1127
V. Ruffato-Ferreira et al. Renewable and Sustainable Energy Reviews 72 (2017) 1124–1137

Water Balance Variation per Time Slice - Eta- Water Balance Variation per Time Slice - Eta-
HadGEM 4.5 - Amazon River Basin HadGEM 8.5 - Amazon River Basin
7 7
6 6
5 5
4 4
3 3
Water Balance

Water balance
2 Minimum 2 Minimum
1 1
0 Maximum 0 Maximum
-1 1961_1990 2011_2040 2041_2070 2071_2100 1961_1990 2011_2040 2041_2070 2071_2100
Average -1
Average
-2 -2
-3 -3
-4 Time period -4 Time period
-5 -5
-6 -6
-7 -7

Fig. 3. Amazon Basin. Variations of minimum, mean and maximum values throughout the present (1961–1990) and future (2010–2040, 2041–2070 and 2071–2100) periods, using
the Eta-HadGEM2 model, a) scenario RCP 4.5 and b) scenario RCP 8.5..

Table 1 3. Results and discussion

Statistical results for the trend of the average water balance of the Amazon River Basin.
3.1. Hydroelectric power plants
Scenario RCP 4.5
Mann-Kendall 0,7341
Sen's Bend −0,0927 Fig. 7 shows the result of the Water Balance index calculated in this
Scenario RCP 8.5 study, distributed throughout the entire territory of Brazil with the
Mann-Kendall 0,0894 mapping technique we employed. The historical time period corre-
Sen's Bend −0,247
sponding to the years 1961–1990 and the three future time periods

Water Balance Variation per Time Slice - Eta- Water Balance Variation per Time Slice - Eta-
HadGEM 4.5 - North and Northeast Atlantic HadGEM 8.5 - North and Northeast Atlantic
7 Basin Basin
6 7
5 6
4 5
3 4
2 3
Water Balance

2
Water balance

1 Minimum
0 1 Minimum
Maximum 0
-1 1961_1990 2011_2040 2041_2070 2071_2100 Maximum
-1 1961_1990 2011_2040 2041_2070 2071_2100
-2 Average
Average
-2
-3
-3
-4
-4
-5 Time period Time period
-5
-6 -6
-7 -7

Fig. 4. North and Northeast Atlantic Basin. Variations of minimum, mean and maximum values throughout the present (1961–1990) and future (2010–2040, 2041–2070 and 2071–
2100) periods, using the Eta-HadGEM2 model, a) scenario RCP 4.5 and b) scenario RCP 8.5.

Table 2 (2010–2040, 2041–2070 and 2071–2100) are presented and calcu-

Statistical results for the trend of the average water balance of the North and Northeast lated through the Eta-HadGEM2 model in two scenarios, RCP 4.5
Atlantic Basin.
(more optimist) and RCP 8.5 (more pessimist).
Scenario RCP 4.5 In general, an increasing trend in areas with WB indexes that are
Mann-Kendall 0,3082 negative (indicating a water deficit) or close to zero is seen in both
Sen's Bend −0,0497 scenarios. The expansion of the area with a water deficit is larger for the
Scenario RCP 8.5 pessimistic scenario. For the projected periods, an increasing trend in
Mann-Kendall 0,0894
Sen's Bend −0,1057
water surplus is observed only for watersheds located in the south of
the country (Paraná-Paraguay, South and Southeast Atlantic, and
Uruguay).
missing values, it is more rigorous than the usual regression curve and A higher water surplus implies a larger hydroelectric power
provides a more realistic measure of trends in a time series. potential for the watershed. Hence, in general, the WB index calculated
through the climate change scenarios modeled by INPE shows a
decreasing trend in hydroelectric power potential for watersheds

1128
V. Ruffato-Ferreira et al. Renewable and Sustainable Energy Reviews 72 (2017) 1124–1137

Water Balance Variation per Time Slice - Eta- Water Balance Variation per Time Slice - Eta-
HadGEM 4.5 - Tocantins-Araguaia Basin HadGEM 8.5 - Tocantins-Araguaia Basin
6
7
5
6
5 4
4 3
3 2

Water balance
2 1
Water balance

1 Minimum Minimum
0
0 Maximum 1961_1990 2011_2040 2041_2070 2071_2100 Maximum
1961_1990 2011_2040 2041_2070 2071_2100 -1
-1 Average Average
-2 -2
-3 -3
-4 Time period -4
Time period
-5 -5
-6 -6
-7 -7

Fig. 5. Tocantins-Araguaia Basin. Variations of minimum, average and maximum values throughout the present (1961–1990) and future (2010–2040, 2041–2070 and 2071–2100)
periods, using the Eta-HadGEM2 model, a) scenario RCP 4.5 and b) scenario RCP 8.5.

Table 3 located in the North and Center of the country and an increasing trend
Statistical results for the trend of the average water balance of the Tocantins-Araguaia in the potential for watersheds located in the south of the country
Basin.
(Fig. 2).
Scenario RCP 4.5
Mann-Kendall 0,7341 3.1.1. Amazon River Basin
SEN's Bend −0,0506 In the Amazon River Basin, a slightly decreasing trend in the
Scenario RCP 8.5
average water balance is observed in both scenarios, significant in
Mann-Kendall 0,3082
SEN's Bend −0,0538 scenario RCP 8.5. An increasing trend in areas with water balances that
are negative or close to zero is observed for both scenarios (RCP 4.5

Water Balance Variation per Time Slice - Eta- Water Balance Variation per Time Slice - Eta-
HadGEM 4.5 - São Francisco River Basin HadGEM 8.5 - São Francisco River Basin
7 7
6 6
5 5
4 4
3 3
Water Balance

2 2
Water balance

1 Minimum
1 Minimum
0 Maximum 0
1961_1990 2011_2040 2041_2070 2071_2100 Maximum
-1 -1 1961_1990 2011_2040 2041_2070 2071_2100
Average
-2 Average
-2
-3 -3
-4 -4
-5 -5
-6 -6
-7 Time period -7 Time period

Fig. 6. São Francisco River Basin. Variations of minimum, mean and maximum values throughout the present (1961–1990) and future (2010–2040, 2041–2070 and 2071–2100)
periods, using the Eta-HadGEM2 model, a) scenario RCP 4.5 and b) scenario RCP 8.5.

Table 4 and 8.5). This trend is significant in scenario RCP 8.5, which can be
Statistical results for the trend of the average water balance of the São Francisco River observed in both Graph 3 and Sen's Bend (Fig. 3 and Table 1). A
Basin.
decreasing trend is seen for the maximum values when minimum
Scenario RCP 4.5 values increase (becoming more negative, thus indicating a larger water
Mann-Kendall 0,7341 deficit). As for power generation, a decreasing trend in the hydro-
SEN's Bend 0 electric power potential is observed for the basin.
Scenario RCP 8.5
Mann-Kendall 0,3082
SEN's Bend −0,0195 3.1.2. North and Northeast Atlantic Basin
The average water balance of the North and Northeast Atlantic
Basin decreases in future years in comparison with the present. The
value of the mean decreases very slightly in scenario RCP 4.5, but a
more significant reduction is observed in scenario RCP 8.5, which

1129
V. Ruffato-Ferreira et al. Renewable and Sustainable Energy Reviews 72 (2017) 1124–1137

Water Balance Variation per Time Slice - Eta- Water Balance Variation per Time Slice -
HadGEM 4.5 - East Atlantic Basin Eta-HadGEM 8.5 - East Atlantic Basin
7 7
6 6
5 5
4 4
3 3
2
Water balance

Water balance
2
Minimum Minimum
1 1
0 Maximum 0 Maximum
-1 1961_1990 2011_2040 2041_2070 2071_2100 -1 1961_1990 2011_2040 2041_2070 2071_2100
Average Average
-2 -2
-3 -3
-4 -4 Time period
-5
Time period -5
-6 -6
-7 -7

Fig. 7. East Atlantic Basin. Variations of minimum, mean and maximum values throughout the present (1961–1990) and future (2010–2040, 2041–2070 and 2071–2100) periods,
using the Eta-HadGEM2 model, a) scenario RCP 4.5 and b) scenario RCP 8.5.

Table 5 Table 9
Statistical results for the trend of the average water balance of the East Atlantic Basin. Statistical analysis of the wind speed at 100 m for the Brazilian regions, in scenario RCP
4.5.
Scenario RCP 4.5
Mann-Kendall 0,7341 Statistical analysis of the Regional Mean Value of the Aridity Indexa
SEN's Bend −0,002
Scenario RCP 8.5 Region Mann-Kendall SEN Curve
Mann-Kendall 0,0894
SEN's Bend −0,0476 Northeast 0,7341 +0,082 m s−1 (per period)
North 0,7341 +0,014 m s−1 (per period)
Midwest 0,7341 −0,008 m s−1 (per period)
Southeast 0,7341 +0,077 m s−1 (per period)
Table 6
South 0,7341 +0,069 m s−1 (per period)
Statistical results for the trend of the average water balance of the Basin of the Rivers
Paraná and Paraguay. a
Analysis of the mean value of the indicator for all points modeled by Eta-HadGem2-
ES for the four selected time slices, with frequency of the data of 30 years.
Scenario RCP 4.5
Mann-Kendall 0,7341
SEN's Bend 0,0099 3.1.3. Tocantins-Araguaia Basin
Scenario RCP 8.5 As Fig. 5 shows, the Tocantins-Araguaia Basin shows a slightly
Mann-Kendall 0,7341
increasing trend in the areas with water balance tending to zero for
SEN's Bend −0,0246
both scenarios (RCP 4.5 and 8.5). In both models, a decreasing trend in
maximum values of the water balance is observed, as well as a slight
Table 7 increase of minimal values, indicating that the water balance of the
Statistical results for the trend of the average water balance of the South and basin tends to be closer to zero (Fig. 5).
Southeastern Atlantic Basin. The analysis conducted for the mean water balance (Table 3) did
not show a statistically significant trend for any scenario. Nevertheless,
Scenario RCP 4.5
Mann-Kendall 0,3082 in spite of the positive values of the average water balance, a decreasing
SEN's Bend 0,1482 trend in water availability is observed for the basin in both models. In
Scenario RCP 8.5 scenario RCP 4.5 the number of points with a negative water balance
Mann-Kendall 0,3082 relatively increases in the basin for the first projected time slice (2011–
SEN's Bend 0,1518
2040), but for the following periods, this percentage lowers. In the RCP
8.5 scenario, the percentage of points with a negative water balance
Table 8 increases continuously, except for the last time slice (2071–2100), in
Statistical results for the trend of the average water balance of the Uruguay River Basin. which it decreases. In general, a decreasing trend in the hydroelectric
power potential is observed for the basin, attributed to the projected
Scenario RCP 4.5 climate changes.
Mann-Kendall 0,7341
SEN Curve 0,1997
Scenario RCP 8.5
Mann-Kendall 0,3082 3.1.4. São Francisco River Basin
SEN Curve 0,1927 As Fig. 6 shows, an increasing trend is seen in the São Francisco
River Basin areas with a negative water balance for both scenarios,
sturdier for scenario RCP 8.5.
showed a statistically significant trend (Fig. 4 and Table 2). In general, Fig. 6 also shows an increasing trend in minimum values, indicating
we can state that the hydroelectric power potential tends to decrease in a larger water deficit. For maximum values, we found a decreasing
the basin. trend in water balance, which tends to zero, as well as the mean values.
The analysis carried out for the mean did not show a statistically
significant trend in any of the scenarios. Nevertheless, in scenario RCP

1130
V. Ruffato-Ferreira et al. Renewable and Sustainable Energy Reviews 72 (2017) 1124–1137

Water Balance Variation per Time Slice - Eta- Water Balance Variation per Time Slice - Eta-
HadGEM 4.5 - Basin of the Rivers Paraná and HadGEM 8.5 - Basin of the Rivers Paraná and
Paraguay Paraguay
7 7
6 6
5 5
4 4
3 3
Water balance

Water balance
2 2
1 Minimum 1 Minimum
0 Maximum 0 Maximum
-1 1961_1990 2011_2040 2041_2070 2071_2100 -1 1961_1990 2011_2040 2041_2070 2071_2100
-2 Average Average
-2
-3 -3
-4 -4
-5 Time period -5 Time period
-6 -6
-7 -7

Fig. 8. Basin of the Rivers Paraná and Paraguay. Variations of minimum, mean and maximum values throughout the present (1961–1990) and future (2010–2040, 2041–2070 and
2071–2100) periods, using the Eta-HadGEM2 model, a) Scenario RCP 4.5 and b) Scenario RCP 8.5.

Water Balance Variation per Time Slice - Eta- Water Balance Variation per Time Slice - Eta-
HadGEM 4.5 - South and Southeast Atlantic HadGEM 8.5 - South and Southeast Atlantic
Basin
7 Basin
6 7
6
5
5
4
4
3 3
Water Balance

2
Water balance

2 Minimum
1 Minimum 1
0 0 Maximum
Maximum
-1 1961_1990 2011_2040 2041_2070 2071_2100 -1 1961_1990 2011_2040 2041_2070 2071_2100 Average
-2 Average -2
-3 -3
-4 -4
-5 Time period -5 Time period
-6 -6
-7 -7

Fig. 9. South and Southeastern Atlantic Basin. Variations of minimum, average and maximum values throughout the present (1961–1990) and future (2010–2040, 2041–2070 and
2071–2100) periods, using the model Eta-HadGEM2, a) scenario RCP 4.5 and b) scenario RCP 8.5.

Water Balance Variation per Time Slice - Eta- Water Balance Variation per Time Slice - Eta-
HadGEM 4.5 - Uruguay River Basin HadGEM 8.5 - Uruguay River Basin

7 7
6 6
5 5
4 4
3 3
Water balance
Water balance

2 Minimum 2 Minimum
1 1
Maximum 0 Maximum
0
1961_1990 2011_2040 2041_2070 2071_2100 -1 1961_1990 2011_2040 2041_2070 2071_2100
-1 Average Average
-2 -2
-3 -3
-4 -4
Time period -5 Time period
-5
-6 -6
-7 -7

Fig. 10. Uruguay River Basin. Variations of minimum, average and maximum values throughout the present (1961–1990) and future (2010–2040, 2041–2070 and 2071–2100)
periods, using the Eta-HadGEM2 model, a) scenario RCP 4.5 and b) scenario RCP 8.5.

4.5, the percentage of points with negative water balance increases in period (2071–2100).
the first time slice (2011–2040) with respect to the historical average, The São Francisco River Basin has the highest minimum values
and then it tends to lower. In scenario RCP 8.5, the percentage increase (Fig. 6) and the highest percentage of negative points (Table 4), since it
is constant for the entire projected period, decreasing slightly in the last is located in a semi-arid region. Nevertheless, the basin presents the

1131
V. Ruffato-Ferreira et al.
Fig. 11. Classes of intensity of wind speed at 100 m ( < 3 m s−1, 3–4 m s−1, 4–5 m s−1, 5–6 m s−1, 6–7 m s−1, 7–8 m s−1, > 8 m s−1) for the present period (1961–90) and future periods
(2010–40, 2041–70 and 2071–2100) in scenarios RCP4.5 and 8.5.
most critical conditions in terms of water deficit, and more detailed
studies should be elaborated on the subject.
3.1.5. East Atlantic Basin
As the graphs show, maximum water balance values tend to
decrease in both scenarios (RCP 4.5 and 8.5), especially in the RCP
8.5 scenario. Moreover, minimum values tend to decrease as well. The
means oscillate slightly in the basin, but they are very close to zero
(Fig. 7).
Once again, a slightly increased in water deficit is seen in scenario
RCP 4.5 for the first time slice (2011–2040), becoming weaker in the
following periods. In scenario RCP 8.5, the analysis conducted for the
mean shows a statistically significant trend for scenario RCP 8.5, with
greater negative and zero water balances homogeneity (Fig. 7).
With this analysis, we conclude that on average, the hydroelectric
power potential of the basin will be slightly altered, with stronger
changes observed if the world reaches the emission standards of RCP
8.5 (Tables 5–9).
3.1.6. Basin of the Rivers Paraná and Paraguay
Fig. 8 shows the graphs constructed for the Basin of the Rivers
Paraná and Paraguay. According to the data, most points of the basin
have a positive water balance during the historical period and the
projected periods in both scenarios (RCP 4.5 and 8.5). The minimum
negative water balance values are close to zero. No statistically
significant trend is found in any of the scenarios when analyzing the
mean.
A slightly increased maximum and mean value of the water balance
is expected for both scenarios (Fig. 8), which may be interpreted as a
growth of water availability in the basin and consequently hydroelectric
power potential.
3.1.7. South and Southeast Atlantic Basin
Despite the mean analysis did not find a statistically significant
trend in any of the scenarios, It can be seen a growth in mean values of
the water balance for both scenarios (RCP 4.5 and 8.5), even if there is
an observed decrease in maximum values. The minimum values are
1132
Renewable and Sustainable Energy Reviews 72 (2017) 1124–1137
close to zero, varying from negative values in the historical period to
positive values in the last periods (Fig. 9). The increasing water
availability can influence the increase of the hydroelectric power
potential.
3.1.8. Uruguay River Basin
The Uruguay River Basin had the most similar results to the climate
modeling data of scenarios RCP 4.5 and 8.5. No water balance deficit is
observed within the basin, i.e., there are no areas with negative water
balance (Fig. 10). The analysis conducted for the mean did not find a
statistically significant trend in any of the scenarios. However, the
mean water balance decreased during 2011–2040, stabilizing and
increasing over the following periods in both scenarios (RCP 4.5 and
8.5). An increased maximum and minimum values is noted; thus the
basin may present isolated increases in water availability, which can
influence the increase hydroelectric power potential, but more detailed
studies are required to confirm this hypothesis.
3.2. Wind power
The maps in Fig. 11 were constructed by calculating the total wind
speed for the points of the INPE model, and it shows the distribution of
total wind speeds divided into classes of intensity ( < 3 m s−1, 3–
4 m s−1, 4–5 m s−1, 5–6 m s−1, 6–7 m s−1, 7–8 m s−1, > 8 m s−1). We
present a historical time slice (1961–1990) and three future time slices
(2010–2040, 2041–2070 and 2071–2100) calculated by the Eta-
HadGEM2 model, scenarios RCP 4.5 (optimist) and RCP 8.5 (more
pessimist).
It is worth pointing out in advance that, by using the means of 30
years without considering seasonality, total wind speed is attenuated.
Hence, the increasing trend in wind speed needs to be corrected. With
that aim, we constructed maps of the variations of wind speed
considering seasonality. Figs. 11, 12 and 13 present the maps of the
variations of the means (of 30 years) of the total wind speed for the
periods of 2010–2040, 2041–2070 and 2071–2100 in comparison with
the historical period (1961–1990), for each season of the year
(summer: DJF; fall: MAM; winter: JJA and spring: SON), in scenarios
V. Ruffato-Ferreira et al. Renewable and Sustainable Energy Reviews 72 (2017) 1124–1137

Fig. 12. Wind speed variation at 100 m for future periods (2010–40, 2041–70 and 2071–2100) in comparison with the present (1961–90) and in scenario RCP 4.5 of the regional
model ETA-HadGEM.

1133
V. Ruffato-Ferreira et al. Renewable and Sustainable Energy Reviews 72 (2017) 1124–1137

Fig. 13. Wind speed variation at 100 m for future periods (2010–40, 2041–70 and 2071–2100) in comparison to present (1961–90) and in RCP8.5 scenario of the regional model ETA-
HadGEM.

1134
V. Ruffato-Ferreira et al. Renewable and Sustainable Energy Reviews 72 (2017) 1124–1137

Table 10 2.9% per year over the projection horizon until 2024, compared to an
Statistical analysis of the wind speed at 100 m for the Brazilian regions, in scenario RCP expansion of the Gross Domestic Product (GDP) of 3.2% per year.
8.5.
To match production growth and meet the demand while attending
Statistical analysis of the Regional Mean Value of the Aridity Indexa the commitments made in INDCs, the Brazilian government signals the
expansion of hydro energy (of 28–33% of the energy matrix by 2030),
Region Mann-Kendall SEN Curve as well as wind, biomass and solar participation for domestic use of
non-fossil energy, confirming the importance of renewable energy in
Northeast 0,0894 +0,269 m s−1 (per period)
North 0,0894 +0, 251 m s−1 (per period) the Brazilian strategic planning [25].
Midwest 0,3082 +0,131 m s−1 (per period) Despite the hydroelectric power plants still show great potential to
Southeast 0,0894 +0,322 m s−1 (per period) be explored in the country, it is also foreseen a progressive increase in
South 0,0894 +0,309 m s−1 (per period) the use of other renewable sources in the expansion of the Brazilian
a energy matrix, especially wind and solar energy [4], as can be seen in
Analysis of the mean value of the indicator for all points modeled by Eta-HadGem2-
ES for the four selected time slices, with frequency of the data of 30 years. the projections of the NEP for 2030 (Table 11).
In relation to large hydroelectric production, the Northern Region
Table 11 is distinguished by significant increase in electricity production capa-
Expansion of the electricity supply in the period 2015–2030, by geographic region (MW). city (43,720 W), being the only source of energy supply expansion in
[2], adapted. the region. With a less significant expansion in terms of large hydro-
electric plants in the Southeast/Midwest, there is an expansion
Source North Northeast Southeast/ South Total
Midwest
planning of about 8860 MW, supplemented by the production of small
hydropower plants (4000 MW).
Large hydroelectric 43,720 580 8860 4140 57,300 As foreseen in the NEP [2], the Northeast has notorious little
plants electricity generation from hydroelectric plants, whether small or large,
Small hydroelectric 0 500 4000 1500 6000
totaling an expansion of only 1080 MW in the period. In northeastern
plants
Wind power plants 0 2200 0 1100 3300 case, the expansion of energy production lies with wind farms
(2200 MW), representing 67% of the national total for this source.
Considering the climate change scenarios, it is observed that the
RCP 4.5 and RCP 8.5, respectively. Amazon and Tocantins Araguaia basins, both in the north of Brazil,
For both scenarios, there is an increased wind speed intensity show a statistically significant decrease in the average water balance in
observed for the North, Northeast and South regions, which defines the most intense climate change scenario. However, this region also has
them as potential areas for wind power production. Wind speed the highest water availability in the country, still allowing the expan-
increases more significantly during summer and fall in scenario RCP sion of hydroelectric generation.
4.5, and during spring and summer in scenario RCP 8.5. Areas where Changes in the water balance for the South and Southeast regions
wind speed is projected to decrease are found mainly in the Midwest point to the maintenance or increase of water availability, not conflict-
and in the State of Amazonas, confirming the fact that these areas will ing with the planned expansion of hydroelectric production. However,
continue not being potential producers of wind power. it is emphasized that they should be examined more carefully before
The analyses of the Mann-Kendall test and Sen's Bend (Tables 10 potential conflicts of water use, due to its dense occupation.
and 11) indicate high statistical significance of the indicator for the The little expansion of hydroelectric production in the Northeast
Brazilian regions only in scenario 8.5, except for the Midwest. The suits the climate change scenarios, since this basin is the one that
trend of the RCP 4.5 scenario, on the other hand, was not statistically presents the most negative water balance of the country.
significant. The magnitude of this variation will be higher in the Related to the wind energy, the results point to an increasing of the
Northeast region for scenario RCP 4.5 and in the Southeast region wind speed in the Northeast, which can contribute to electricity
for scenario RCP 8.5. In general, the results indicate an increasing generation. Thus, it is important to conduct studies to estimate the
trend in average wind speed for all regions in Brazil, in both scenarios, real growth of areas with wind speeds suitable for energy production.
except for the Midwest region in scenario RCP 4.5, which suggests a Considering the projection of the wind potential [41] and the results
positive impact on the analysis of this energy source. obtained, the installation of wind farms in this region is a sound choice,
It is worth singling out that, according to Chou et al. [34], the Eta- fitting the government to promote an improvement and expansion of
HadGEM2 model has an overestimated temperature for the region of interconnections of these plants to the NIS.
the State of Amapá, which may correspond to an overestimation of Similarly, climate modeling indicates a rising trend of wind speeds
wind speed in the region due to a consequent increase of pressure. For for the Brazilian north coast, especially in the states of Amapá, Pará
the RCP 8.5 scenario, this trend may cause an even more significant and Roraima, pointing to the need for investment in research on the
error in the overestimation of the potential for wind power generation, production potential of this region and support the implementation of
in comparison with the RCP 4.5 scenario. So far, there were no other new wind farms. Currently, the North does not have wind expansion
limitations of the model described in the literature. planning [2], reiterating the need for a strategic planning linked to
Lastly, for a more accurate assessment of the tendency to use wind climate understanding, since it may be incurring an underestimation of
for power generation, land use projections based on the climate the potential in the Climate Change scenario. The exploration of this
projections of the model used in this study would be required. As wind potential in northern region may represent the completion of
such, it is worth highlighting that this study presents a simplified hydroelectric generation, reducing the social and environmental im-
evaluation of total wind speed. The impacts of terrain roughness on the pacts of large hydroelectric plants [5] and ensuring the protection of
climate projections of wind speeds are extremely important and should the Amazon biome.
be included in future analyses.
4. Conclusions
3.3. Brazilian renewable energy planning face to climate changes
scenarios This study aimed at calculating indicators and indexes to assist in
the analysis of the vulnerability and adaptation of the renewable energy
According to the Energy Research Company [4], it is expected that sector to climate change, focusing specifically on hydroelectricity and
the final energy consumption in Brazil will grow at an average rate of wind power sources, in order to present the importance of incorporat-

1135
V. Ruffato-Ferreira et al.
ing the study of climate change in the Brazilian electric sector planning.
In general, we can conclude that the energy matrix of Brazil has certain
sensitivity to climate variations in the light of two renewable energy
sources analyzed. Understanding the interaction between future cli-
mate and renewable electric power plants is of strategic importance to
the planning of the energy sector, making studies on the subject
essential to increase system resiliency.
The most important result is related to the changing behavior of
environmental variables due to global climate change, affecting hydro
and wind power vulnerability in Brazil. New researches are needed in
order to associate the results presented here with the variation of the
production potential, and thus provide energy allocation infrastructure
scenarios.
Related to the water balance, it was observed an increasing trend in
water scarcity for almost the entire territory of Brazil, especially in its
central region, for the upcoming years (2011–2040), with a small
recovery occurring over the next periods. At the same time, water
availability in the South region of the country will progressively
increase, starting from the current scenario. The most vulnerable water
basin, with the worst scenario of increased water scarcity, is the São
Francisco River in northeastern Brazil.
As for wind power, an increasing trend in total wind speed is
observed, especially in the North, Northeast and South regions,
expanding the country's potential to generate wind power according
to the climatological data projected by the studied model. We highlight
that, without a significant increase in the efficiency of the cost-benefit
relationship of wind power, this energy source will only play a
complementary role to hydroelectric power.
It should be noted the great potential for wind power in the
Northeast of Brazil and the indicated trend by the model of increase
in areas with harnessable wind speed. Investment in wind farms in this
area can be strategic to compensate the possible reduction in the supply
of hydroelectricity with the water availability reduction, while also
stimulating local economic development and diversification, the poor-
est region in the country.
These predictions should be taken in consideration cautiously as a
more correct and robust analysis needs to take into account the
vegetation of biomes since it plays a fundamental role in water balance.
As every reality modeling process, a major source of uncertainty in this
study is related to the uncertainties and inaccuracies of the climate
model employed. Another limitation on climate modeling is linked to
the low level of detail on the projected scenarios of land use and land
cover change, critical to both assessed energy matrix. Thus, all
projections indicated here should be used with caution for the
foundation of public policy and planning.
Aiming to achieve the commitments made by INDCs, the Brazilian
government, according to energy programs formulated for the country
(NEP and DPE), has invested in expanding the use of wind and solar
energy, as a complement to hydropower. Other factor for the increased
planning is the climate change, which have strong evidences of
alterations in the present and future water regimes of the Brazilian
basins. Thus, it is important to conduct studies to better estimate the
wind speed and solar radiation available.
It is worth singling out that this article is based on current
technology; hence, technological improvements can alter the presented
scenario. Lastly, the recommendations of this study are: 1) technology
investments to make wind power turbines more economically profit-
able; 2) the need for a more diverse energy matrix in Brazil, aiming at
lower vulnerability and higher resilience; 3) energy sources in Brazil
need continuous planning, considering the repercussion of new cli-
mate-environmental inputs in the potential for power generation and
consumption demand, so that Brazilian development is not limited by
the (lack of) energy offer; 4) there is a great need of improved regional
climate projection models, mainly in the Amazon region,
1136
Renewable and Sustainable Energy Reviews 72 (2017) 1124–1137
Acknowledgements
The authors acknowledge the support and trust of UNDP via MCTI,
which allowed the preparation of this study in the context of the UNDP
Project BRA/10/G32, essential in the national strategic planning. This
paper is also a contribution of the Brazilian Network on Global Climate
Change Research funded by CNPq Grant Number 550022/2014-7 and
FINEP Grant number 01.13.0353.00. In addition, we acknowledge
INPE and its advisers, the availability to meet the needs and questions
related to the extracted data of climate modeling. Furthermore, the
International Virtual Institute of Global Change (IVIG/COPPE/UFRJ),
for providing the necessary infrastructure to achieve the results.
Finally, the Professor André FP Lucena, PPE/COPPE/UFRJ, whose
explanations about his previous work in the Second National
Communication allowed the continuity of this process.
References
[1] Agência Nacional de Águas (ANA). Conjuntura dos Recursos Hídricos no Brasil:
Encarte especial sobre a crise hídrica. Brasília – DF, Brasil. 〈http://www3.snirh.
gov.br/portal/snirh/snirh-1/conjuntura-dos-recursos-hidricos/crisehidrica2014.
pdf〉; 2015 [accessed 15.05.15].
[2] Empresa de Pesquisa Energética (EPE). Plano Nacional de Energia (PNE) 2030,
〈http://www.epe.gov.br/PNE/Forms/Empreendimento.aspx〉; 2008 [accessed 10.
05.16].
[3] Empresa de Pesquisa Energética (EPE). Balanço Energético Nacional (BEN),
〈https://ben.epe.gov.br/downloads/Relatorio_Final_BEN_2015.pdf〉; 2015
[accessed 10.05.2016].
[4] Empresa de Pesquisa Energética (EPE). Plano Decenal de Energia (PDE), 〈http://
www.epe.gov.br/PDEE/Sum%C3%A1rio%20Executivo%20do%20PDE%202024.
pdf〉 [accessed 10.07.2016].
[5] Lucena AFP, Szklo AS, Schaeffer R, Souza RR, Moreira BS, Borba C, da Costa IVL,
Júnior AOP, da Cunha SHF. The vulnerability of renewable energy to climate
change in Brazil. Energy Policy 2009;37:879–89. http://dx.doi.org/10.1016/j.en-
pol.2008.10.029.
[6] Haylock MR, Peterson TC, Alves LM, Ambrizzi T, et al. Trends in total and extreme
South American rainfall in 1960-2000 and links with sea surface temperature. J
Clim 2005;19:1490–512.
[7] Obregón G, Marengo JA. Mudanças Climáticas Globais e seus Efeitos sobre a
Biodiversidade: Caracterização do Clima no Século XX no Brasil: Tendências de
Chuvas e Temperaturas Médias e Extremas. MMA Relatório 2007;2.
[8] Marengo JA, Rusticucci M, Penalba O, Renom M. An intercomparison of observed
and simulated extreme rainfall and temperature events during the last half of the
twentieth century: Part 2: historical trends. Clim. Change 2010;98:509–29.
[9] Ambrizzi T, Rocha R, Marengo JA, Pisnichenko I, Alves LM. Cenários
Regionalizados de Clima no Brasil para o Século XXI: Projeções de Clima usando
Três Modelos Regionais. Relatório 3, Ministério do Meio Ambiente (MMA), DCBio
Mudanças Climáticas Globais e Efeitos sobre a Biodiversidade – Subprojeto:
Caracterização do Clima Atual e Definição das Alterações Climáticas para o
Território Brasileiro ao longo do Século XXI; 2007.
[10] Marengo JA, Jones R, Alves LM, Valverde M. Future change of temperature and
precipitation extremes in South America as derived from the PRECIS regional
climate modeling system. Int J Climatol 2009;29:2241–55.
[11] Marengo JA, Chou SC, Kay G, Alves LM, Pesquero JF, Soares WR, Santos DC, Lyra
AA, Sueiro G, Betts R, Chagas DJ, Gomes JL, Bustamante JF, Tavares P.
Development of regional future climate change scenarios in South America using
the Eta CPTEC/HadCM3 climate change projections: climatology and regional
analyses for the Amazon, São Francisco and the Parana River Basins. Clim Dyn
2012;38:1829–48.
[12] International Energy Agency (IEA). Key World Energy Statistics, 〈http://www.iea.
org/textbase/nppdf/free/2011/key_world_energy_stats.pdf〉; 2011 [accessed 21.
09.14].
[13] Empresa de Pesquisa Energética (EPE). Ten-year energy expansion plan 2024.
Brasília – DJ, Brasil, 〈http://www.epe.gov.br/PNE/Forms/Empreendimento.
aspx〉; 2008 [accessed 10.05.16].
[14] Conselho Empresarial Brasileiro para o Desenvolvimento Sustentável (CEBDS).
Estudo sobre adaptação e vulnerabilidade à variabilidade climática: casos do setor
elétrico brasileiro. Rio de Janeiro, Brazil; 2013.
[15] Schaeffer R, Szklo AS, Lucena AFP, Souza RR, BSMC Borba, IVL Costa, Pereira Jr
AO, SHF Cunha. Mudanças Climáticas e Segurança Energética no Brasil. Rio de
Janeiro: Nova Brasileira; 2008. p. 67.
[16] Milly PCD, Dunne KA, Vecchia AV. Global pattern of trends in streamflow and
water availability in a changing climate. Nature 2005;438:347–50.
[17] Bravo JM, Collischonn W, Silva, BC, Rodriguez D, Fan FM, Quiroz K, Pontes PRM.
Impactos nas Afluências das Principais Bacias. Efeitos das Mudanças Climáticas na
Geração de Energia Elétrica, AES Eletropaulo e AES Tietê, Capítulo 6; 2014.
[18] Agência Nacional de Energia Elétrica (ANEEL). BIG ANEEL database 〈http://www.
aneel.gov.br/aplicacoes/capacidadebrasil/capacidadebrasil.cfm〉 [accessed 12.05.
16].
[19] Grubb MJ, Meyer NI. Wind energy: resources, systems and regional strategies,
V. Ruffato-Ferreira et al.
1993. In: Agência Nacional de Energia Elétrica (ANEEL), Atlas de Energia Elétrica;
2005, p. 243.
[20] de Jong P, Sánchez AS, Esquerre K, Kalid RA, Torres EA. Solar and wind energy
production in relation to the electricity load curve and hydroelectricity in the
northeast region of Brazil. Renew Sustain Energy Rev 2013;23:526–35.
[21] Sánchez AS, Torres EA, Kalid RA. Renewable energy generation for the rural
electrification of isolated communities in the Amazon Region. Renew Sustain
Energy Rev 2015;49:278–90.
[22] Agência Nacional de Energia Elétrica – ANEEL. Atlas de Energia Elétrica, 2nd ed.;
2005, p. 243.
[23] Pereira EB, Martins FR, Pes MP, Segundo EIC, Lyra AA. The impacts of global
climate changes on the wind power density in Brazil. Renew Energy
2013;49:107–10.
[24] Lucena AFP, Szklo AS, Schaeffer R, Dutra RM. The vulnerability of wind power to
climate change in Brazil. Renew Energy 2010;35:904–12.
[25] Brazil. Brasil intended Nationally Determined Contribution – iNDC. Brasília – DF,
Brasil. 〈http://www.itamaraty.gov.br/images/ed_desenvsust/BRASIL-iNDC-
portugues.pdf〉; 2015 [accessed 10.05.16].
[26] Moss RH, Edmonds JA, Hibbard KA, Manning MR, Rose SK, van Vuuren DP,
Carter TR, Emori S, Kainuma M, Kram T, et al. The next generation of scenarios for
climate change research and assessment. Nature 2010;463:747–56.
[27] Intergovernmental Panel on Climate Change - IPCC. Climate Change 2013: The
Physical Science Basis. Contribution of Working Group I to the Fifth Assessment
Report of the IPCC [Stocker TF , Qin D, Plattner G-K, Tignor M, Allen SK,
Boschung J, Nauels A, Xia Y, Bex V, Midgley PM(Eds.)]. , Cambridge, United
Kingdom and New York, NY, USA: Cambridge University Press, p. 1535. 〈http://
dx.doi.org/10.1017/CBO9781107415324〉.
[28] Van Vuuren DP, Edmonds J, Kainuma M, et al. The representative concentration
pathways: an overview. Clim Change 2011;109:5–31.
[29] Thomson AM, Calvin KV, Smith SJ, et al. RCP4.5: a pathway for stabilization of
radiative forcing by 2100. Clim Change 2011;109:77–94.
[30] Intergovernmental Panel on Climate Change - IPCC. Climate change 2014: impacts,
adaptation, and vulnerability. Part A: global and sectoral aspects. Contribution of
working group II to the fifth assessment report of the intergovernmental panel on
climate change [Field CB, Barros VR, Dokken DJ et al. (eds.)]. Cambridge, United
Kingdom and New York: Cambridge University Press; 2014, p. 1132.
[31] Riahi K, Krey V, Rao S, et al. RCP-8.5: exploring the consequence of high emission
trajectories. Clim Change 2011;109:33–57.
1137
Renewable and Sustainable Energy Reviews 72 (2017) 1124–1137
[32] Black TL. The new NMC mesoscale Eta Model: description and forecast examples.
Weather Forecast 1994;9:265–78.
[33] Mesingeret F, et al. I. An upgraded version of the Eta Model. Meteorology and
Atmospheric Physics; 2012.
[34] Intergovernmental Panel on Climate Change – IPCC. Climate Change 2014:
Synthesis Report. Contribution of working groups I, II and III to the fifth
assessment report of the intergovernmental panel on climate change [Core Writing
Team, Pachauri RK, Meyer LA (eds.)]. Geneva, Switzerland: IPCC; 2014, p. 151.
[35] Chou S, Lyra A, Mourão C, Dereczynski C, Pilotto I, Gomes J, Bustamante J,
Tavares P, Silva A, Rodrigues D, Campos D, Chagas D, Sueiro G, Siqueira G, Nobre
P, Marengo J. Evaluation of the Eta simulations nested in three global climate
models. Am J Clim Change 2014;3:438–54.
[36] Chou , et al. Assessment of climate change over South America under RCP 4.5 and
8.5 downscaling scenarios. Am J Clim Change 2014;3:512–27.
[37] Intergovernmental Panel on Climate Change – IPCC. Climate Change 2013: Fifth
Assessment report of the intergovernmental panel on climate change. Stocker TF,
Qin D, Plattner GK, Tignor M, Allen SK, Boschung J, Nauels A, Xia Y, Bex V,
Midgley PM, Eds., Cambridge, United Kingdom and New York: Cambridge
University Press, Cambridge, 1535 p., 2013.
[38] Souza OTL de. Desenvolvimento de um modelo fuzzy para determinação do calor
latente com aplicação em sistemas de irrigação. UNESP, Aleph xi; 2004 p.113.
[39] Food and Agriculture Organization of the United Nations. Crop evapotranspiration:
guidelines for computing crop water requirements, FAO irrigation and drainage
paper. Food and Agriculture Organization of the United Nations, Rome; 1998.
[40] Agência Nacional de Energia Elétrica - ANEEL. Bacias Hidrográficas do Brasil
segundo ao Dnaee. 〈http://www.aneel.gov.br/area.cfm?id_area=104〉 [accessed
22.08.14].
[41] Amarante OAC, Zack J, Brower M, Sá AL. Atlas do Potencial Eólico Brasileiro.
Centro de Referência para Energia Solar e Eólica Sergio de Salvo Brito – CRESESB;
2001.
[42] Sneyers R. Sur L’analyse Statistique des Series Dóbservations. Gênevè:
Organisation Méteorologique Mondial. Gênevè: Organisation Méteorologique
Mondial; 1975. p. 192.
[43] Goossens C, Berger A. Annual and seasonal climatic variations over the Northern
Hemisphere and Europe during the last century. Ann Geophys 1986;4:385–400.
[44] Sen PK. Estimates of the regression coefficient based on Kendall's Tau. J Am Stat
Assoc 1968(324):1379–89.