Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

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

Comparison of 100% renewable energy system scenarios with a focus on

flexibility and cost
Wesley Deason
International Atomic Energy Agency, PO Box 100, 1400 Vienna, Austria

A R T I C L E I N F O A B S T R A C T

Keywords: In this study, a particular class of energy system scenarios assuming 100% renewable energy (RE) are selected
Energy system flexibility and compared. The purpose is to compare the relative characteristics of the scenarios, focusing on the amount
Energy system modeling and types of flexibility used and the predicted electricity generation cost. The approach included the screening of
High-penetration renewable energy systems 45 studies (some of which contained multiple scenarios) to down-select to 8 studies, which used high-fidelity
short-term energy system models. The compared scenarios applied many strategies for providing flexibility,
which was found to be a crucial and significant component of a 100% RE system. Accordingly, only two of the
fifteen examined scenarios contained a variable generation fraction greater than 65% of total scenario capacity
(73% and 69%). Predicted electricity costs varied drastically in all regions. Some studies provided multiple
scenarios (including business-as-usual scenarios and costs) that could be compared. In 3 out of the 4 studies
including a business-as-usual cost, the 100% RE scenarios were found to be between 41% and 104% more
expensive. The fourth study showed electricity costs to be the same for a 100% RE system and business-as-usual
system, and the 100% RE system to be 62% cheaper when externalities are included.

1. Introduction been selected and compared.1 Particular focus has been given in de-
termining the amount and types of flexibility used in the scenarios and
In high-penetration renewable energy (RE) systems, variable re- the predicted electricity generation cost. Though the ability for the
newable sources of electricity generation (e.g. wind and solar power) scenarios to provide time-dependent electric grid load balancing has
are used extensively. To achieve maximum use of these generation been ensured through the use of a short-term energy system model,
sources, the system must compensate in some way the inherent varia- their ability to provide geographical load balancing and other ancillary
bility of their power output. The energy system's ability to compensate services has not. Provision of all ancillary services to the electric grid
for this variability is commonly referred to as flexibility. Renewable (e.g. frequency control, voltage control, system restoration, and system
sources of flexibility can be added to the energy system through many control) remains necessary for its safe and reliable operation [2].
methods, such as energy generation (e.g. hydroelectric, biofuels, and
concentrating solar thermal power with storage - CSP), grid inter- 1.1. Outline
connections, energy storage (e.g. thermal storage, compressed air,
pumped storage, batteries, or synthetic fuels), and flexible demand. The Section 2 reviews the literature, focusing on four previous reviews
amount of flexibility required in a high-penetration RE system can be of high-penetration RE system studies. Section 3 presents this study's
higher than expected at first glance. Energy generation troughs from methodology. Section 4 presents a wide selection of the high-penetra-
variable sources may align with energy demand peaks for an extended tion RE system studies that have been published and describes those
period of time, leading to a shortage of energy. Proper validation picked for in-depth analysis. Section 5 presents analysis results, with
through use of a short-term energy system model can capture this subsections devoted to scenario demand, scenario capacity, and sce-
phenomenon, ensuring the proper amount of flexibility is designed into nario cost. Section 6 summarizes the study's conclusions.
the system [1].
In this study, several 100% RE scenarios for various regions have

E-mail address: [email protected].

1
This work was inspired by a presentation given by Ben Heard, a PhD student at the University of Adelaide, Australia at the World Nuclear Association Symposium 2015 in London,
UK. The presentation consisted of a review of several different decarbonization studies. Results of that presentation are not discussed in the present section because it is not available in
public literature.

https://doi.org/10.1016/j.rser.2017.10.026
Received 23 February 2017; Received in revised form 4 August 2017; Accepted 26 October 2017
1364-0321/ © 2017 Published by Elsevier Ltd.

Please cite this article as: Deason, W., Renewable and Sustainable Energy Reviews (2017), https://doi.org/10.1016/j.rser.2017.10.026
W. Deason Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx

2. Literature review needed to produce scenarios verified with economically-based capacity

expansion and dispatch models.
The publications discussed below reviewed studies of high-pene- Schmid et al. reviewed 10 climate change mitigation scenarios for
tration RE scenarios. The studied scenarios ranged from general sce- Germany [16]. The authors reviewed all aspects of the scenarios in
narios—which consider the capability for renewable resources to pro- order to determine and compare their overarching strategies. It found
vide energy to different regions of the world over a long time their main strategies were to increase domestic renewable electricity
period—to specific scenarios—which examine the functioning of a production, to reduce electricity demand, and to rely on imports of
single region's electricity grid over a short and highly-discretized time renewable electricity from neighbouring countries. Though the review
period. discussed cost numbers found in several of the scenarios, the lack of a
Loftus et al. compared 17 global decarbonization scenarios with the common methodology or goal prevented a useful comparison. Through
goal of understanding their feasibility [3]. To do so, the authors ex- discussion of the underlying assumptions of the scenarios and by
amined the scenarios’ change in energy intensity, change in carbon comparing those assumptions with the scenario results, the authors
intensity, and low-carbon technology growth rates relative to historical conclude that there are still barriers in place related to technology
data and future projections. Notable results were that all scenarios development and implementation preventing the strategies suggested
featured a reduction in future per-GDP energy use between 2010 and in the scenarios from being carried out.
2050 compared to historically observed rates. 14 of the 17 scenarios The present work adds to the work done in the publications dis-
featured low-carbon technology growth rates similar to historical ex- cussed above by focusing specifically on studies verified using short-
perience. The 3 outlier scenarios ( [4,5]/[6], and [7]) required much term (or “dispatch”) models. Additionally, the present work adds value
higher deployment rates to meet their decarbonization targets due to by investigating the role of flexibility in proposed 100% RE systems.
the a priori decision to exclude some low-carbon technologies (e.g.
nuclear, carbon capture and storage, and biomass). Economic costs 3. Methodology
from the scenarios were reviewed, with all except one showing an in-
cremental investment in the range of 1% of global GDP (20–50% over The first step taken in the present analysis was the selection of re-
the baseline scenario's investments). It is also noted that the 17 ex- levant energy system studies. To determine study relevancy, four main
amined scenarios featured little details on how transportation and in- criteria were used: inclusion of a 100% RE scenario; verification of the
dustrial sectors could be decarbonized, and how key constraints (e.g. 100% RE scenario using a short-term model; a suitable geographical
technical, economic, infrastructural, and societal) could be overcome. scope enabling comparison with obtainable energy statistics; and the
Reedman looked at 23 high-penetration RE studies in the context of reporting of information related to the assumed power capacities for all
Australia and discussed their characteristics [8]. In particular, it energy generation, storage, grid interconnection, and demand flex-
documented the studies’ proposed resource use and technologies, ibility technologies. Only 100% RE scenarios were considered in order
methods, assumptions, and results – including the estimated generation, to limit complexity of comparison results and to provide a better
generation capacity, and overall cost. The review examined studies baseline for comparing scenarios against. Since the purpose of the
focusing on RE penetration into electricity grids of similar size to or present study is not to determine the feasibility of the studied scenarios,
larger than the Australian National Electricity Market's grid. Reedman's validation by a short-term energy system model is assumed to be suf-
study was commissioned by the Australian Energy Market Operator to ficient enough to ensure feasibility.
aid in the development of their own 100% RE study [9]. Notable con- Next, as much information as possible was gathered from the sce-
clusions in the review were that the studied scenarios collectively narios about assumed energy generation and demand, power capacities,
identified challenges due to resource availability; the integration of RE and costs. The information was then categorized and compared. The
into the grid; and social, environmental, and institutional constraints. categories, which are listed in Table 1, were designed to best encompass
However, the studies found these challenges were ultimately sur- the role each technology plays in the system.
mountable. Similar to Loftus et al., Reedman compared scenario costs to The principal category, ‘Total Scenario Capacity’, groups together
a baseline scenario [3]. The results showed a 10–50% increase relative all types of capacity capable of adding or flexibly removing energy from
to the baseline retail electricity price chosen by Reedman from a 2012 the system. This categorization is counter-intuitive to the way energy
NREL study [10]. The review also summarized two studies ( [11] and systems have been previously described and requires some explanation.
[12]) on the transformation of the Australian electricity system to 100% One main attribute all technologies within total scenario capacity share
RE, estimating total costs of 317 billion 2010AUD (~356 billion is a real cost associated with their construction or utilization. The ca-
2012USD) and 370 billion 2010AUD (~415 Billion 2012USD) respec- pacities also have the same unit, power. By adding or removing energy
tively. Reedman also identified shortcomings in the studied scenarios from the system, they can ensure the distribution and transmission in-
related to the lack of the following: explicit modeling of the electric frastructure is efficiently loaded (not over or underutilized). The tech-
grid, market conditions and demand behaviour (demand side partici- nologies work in sync to ensure the efficient and reliable operation of
pation or vehicle-to-grid); analysis on the effects of institutional and the 100% RE system. Lastly, by tracking the total scenario capacity, it is
government policies; and analysis on how the proposed systems affect desired that the individual capacities can be compared side by side and
economic performance statistics regarding jobs and growth. with present day technology capacities. Rather than just replacing
Cochran et al. reviewed 12 high penetration RE scenarios, com- current capacity with new generation capacity, flexible demand and
paring their methodologies, assumptions, and results [13]. Similar to storage are being brought online as well and should accordingly be
Reedman, the analysed scenarios consisted of a mix of global and re- included in comparisons.
gional scenarios [8]. The authors compared costs calculated by the The subordinate categories, ‘Variable Generation Capacity’ and
scenarios, finding that some of the studies showed incremental cost ‘Capacity for Flexibility’ split the technologies according to their ability
increases relative to a reference scenario, one [14] showed “competi- to provide flexibility. Variable generation capacity is composed of solar
tive” costs; and another [15] showed RE systems would be cheaper.2 photovoltaic (solar PV), wind, and ocean technologies. These electricity
The review ultimately concludes that RE technologies vary con- generation technologies produce usable energy at variable rates and
siderably on a regional and global scale, and that more research is any energy collected by these technologies must be used immediately or
stored in an external storage facility. The variable generation capacity
category highlights the importance of variable generating technologies
2
A further description of these costs can be found Table 5 of Cochran et al.’s review and helps in understanding their penetration into the overall system.
[13]. The other category, ‘Capacity for Flexibility’, groups together

2
W. Deason Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx

Table 1 Together these categories cover all flexible energy supply compo-
Categorization of technologies in a 100% RE system. nents used in the reviewed scenarios.
Total scenario Variable generation capacity Solar photovoltaic
capacity On shore wind 4. Studies selected for review
Off shore wind
Ocean (tidal and wave) A catalogue of high-penetration renewable energy studies is shown
Capacity for Flexible Concentrating solar in Table 2. The studies were examined using the criteria discussed in
Flexibility generation thermal power with
capacity storage
the previous section. They cover a wide range of regions and represent
Solar heating with storage many different strategies for use of large amounts of RE in electricity
Hydroelectric (run-of-the- and thermal energy generation. The selection of the studies was done
river and reservoir) based upon the four rightmost columns in Table 2.
Geothermal
Table 3 lists the final selection of studies along with characteristics
Biofuels
Transmission Grid interconnection of the energy system model they build on. To allow comparison of
capacity scenarios with those closest to them in scale, the scenarios have been
Storage capacity Non-CSP thermal storage arranged according to the 2013 electricity demand of the country they
Compressed air energy target [60]. This arrangement of the scenarios is used throughout the
storage (CAES)
Hydroelectric with
rest of the article. Some of the studies contain multiple scenarios, and
pumped storage cover a large geographical range. Most of the models used to evaluate
Hydrogen electricity the scenarios use a dispatch interval of 1 h and a simulation length of 1
generation capacity year (both historic and future years up until 2050 are being con-
Hydrogen heat generation
sidered). This configuration is useful for connecting to publically
capacity
Battery capacity available electricity demand statistics and variable renewable resource
Flexible demand Demand side participation availability statistics. If more discrete information is available, then the
Electric vehicles dispatch interval can be decreased accordingly. A distinctive study is
Hydrogen electrolysis that of Jacobson et al.’s with a dispatch interval of thirty seconds and a
Reducible Large industrial users with
demand pre-arrangements for
simulation length of six years, which uses a computer model to generate
energy reduction energy demand and RE resource availabilities [51]. Another distinctive
study by Krakowski et al. models the energy supply during re-
presentative periods of a common day, week, and year [32]. By redu-
technologies that produce or use energy at controllable rates. These cing the number of time steps needed to analyse a year, the authors
technologies can provide flexibility to a 100% RE system, enabling could allow the model to determine the evolution of the energy system
more economic utilization of energy. By grouping them together, it can over a period of 40 years. The study specifically addresses the possi-
be seen that some form of capacity for flexibility is necessary in the bility of a period with high energy demand and low wind and solar
reviewed 100% RE scenarios. Capacity for flexibility can be broken resource availability by evaluating a specific period in January-Feb-
down further based upon how each specific category of technology ruary with low wind and solar availability.
operated with respect to the system.
5. Scenario analysis results
• Flexible generation capacity is composed of CSP with storage, solar
heating with storage, hydroelectric (run-of-the-river and reservoir), 5.1. Scenario demand
geothermal, and biofuels. These generation technologies generally
feature some amount of internal or inherent storage (where energy Determination of energy demand is crucial to developing the system
does not have to be used immediately) or tap into a consistently capacity needed in a future energy scenario. It is highly dependent on
available energy or fuel reservoir (as in geothermal or run-of-the- assumptions made on the evolution of energy-dependent sectors
river). (transportation, industry, heating, etc.). Adoption of new technologies
• Transmission capacity is composed of electricity transmission ca- and lifestyle changes may drastically affect the amount and types of
pacity that allows a region to trade electricity with neighbors. energy needed. Table 4 shows the change in the demand in the various
• Storage capacity is composed of all storage technologies used in the scenarios compared to final electricity and energy consumption values
system not integrated directly or inherently within a single flexible from 2013. Jacobson et al. is distinctive from the other studies due to its
generation technology. The storage technologies considered in the scope, which assumes that RE produces all energy for energy-dependent
present work are: non-CSP thermal storage, compressed air energy sectors in the United States [51]. The other scenarios are not as ex-
storage (CAES), hydroelectric with pumped storage, hydrogen tensive. Despite energy saving and efficiency measures, some scenarios
electricity generation capacity, hydrogen heat generation capacity, forecast electricity demand to stay constant or increase over the long
and battery capacity. Energy for these storage technologies can term. Other scenarios simply use unaltered present day electricity de-
generally be stored from any variable or flexible generation tech- mand.
nology or from a combination of technologies (i.e. energy from an
electricity or heat market). 5.2. Scenario capacity
• Flexible demand is defined as energy demand which can be shifted
within a predefined period of time (usually less than 24 h) to use The amount of flexibility needed varies based on the geographical
energy as soon as it's available or to delay if it is not. An example of and regional constraints applied to each scenario. To aid in seeing the
flexible demand is electricity used to supply energy to electric ve- capacity for flexibility needed in all the scenarios, all categories except
hicles (EVs) or for hydrogen electrolysis. variable generation capacity have been grouped together in Fig. 1
• Reducible demand is demand not required to be met within any time under the label ‘Capacity for Flexibility’. Variable generation capacity is
period. An example of reducible demand is energy supplied to a shown alongside the capacity for flexibility, both of which when
manufacturing plant, which may be willing to reduce output in summed together create the total system capacity (as defined in the
exchange for payment according to a pre-arranged agreement. methodology section). In Fig. 1, only two of the fifteen scenarios feature
a variable generation fraction greater than 65% of the total system

3
W. Deason Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx

Table 2
Catalogue of high-penetration RE studies. An “X” denotes a fulfilled study criterion. Studies selected for in-depth review are bolded.

Region Year Main article author or report institution 100% RE Uses short-term Country-Level Contains all capacity-specific
model scope dataa

Australia 2010 Beyond Zero Emissions [12] X X X

Australia 2012 Trainer [17] X X
Australia - NEM Gridb 2012 Elliston et al. [18] X X See notec X
Australia - NEM Gridb 2013 AEMO [9] X X See notec X
Australia - NEM Gridb 2016 Huva et al. [19] X X
Brazil 2014 Schmidt et al. [20] X X
Croatia - Mljet 2009 Krajačić et al. [21] X X X
Denmark 2008 Lund and Mathiesen [22] X X X
Denmark 2015 Mathiesen et al. [23] X X X
Denmark 2015 Østergaard et al. [24] X X X
Europe 2010 EREC [25] X X X
Europe 2010 Greenpeace and EREC [26] X X
Europe 2013 Trainer [27] X X
Europe 2016 Bertsch et al. [28] X X X
Europe - Northern Countries 2008 Sørenson [29] X X X
Europe - Southeast Countries 2016 Dominković et al. [30] X X X
Europe and North Africa 2011 energynautics GmbH [31] X X X
France 2016 Krakowski et al. [32] X X X X
France - Reunion Island 2015 Drouineau et al. [33] X X
Germany 2011 SRU [34] X X X X
Germany 2012 DLR [35] X X X
Great Britain 2013 Centre for Alternative Technology [36] X X
India - Kerala 2013 WWF-India and WISE [37]
Ireland 2011 Connolly et al. [38] X X X
Japan 2003 ISUSI [39] X X X
Japan 2012 Esteban et al. [40] X X X
Macedonia 2012 Ćosić et al. [41] X X X
New Zealand 2013 Mason et al. [42] X X X X
Portugal 2010 Krajačić et al. [43] X X X X
Portugal 2013 Fernandes and Ferreira [44] X X X X
South Africa 2011 Greenpeace [45] X X
Spain - Canary Islands 2015 Greenpeace [46] X X
Spain - Catalonia 2007 Fundació Terra, iSuSI, Ecoserveis [47] X X X
Sweden 2011 IVL [48] X X
United Kingdom 2015 Pfenninger and Keirstead [49] X X X
United States 2012 NREL [50] X X X
United States 2015 Jacobson et al. [51] X X X X
United States 2015 Jacobson et al. [52] X X X
United States 2008 Fthenakis et al. [53] X X
United States - PJM Grid 2012 Budischak et al. [54] X X
United States - WECC Grid 2011 Short and Diakov [55] X X X
World 2011 WWF [4] X X
World 2013 Trainer [56] X X
World 2015 Greenpeace, GWEC, SolarPowerEurope X X X
[57]
World 2016 Jacobson et al. [58] X X X

a
Though some of the scenarios met most or all of the selection criteria, some lack key information on the capacities of some system components. Without this information, it was not
possible to compile information needed for in-depth analysis.
b
The National Electricity Market (NEM) is the name for the wholesale electricity market covering the eastern and southern coasts of Australia. It supplies around 80% of Australia's
electricity consumption [59].
c
An exception was made for the two Australian NEM Grid studies because present day capacity information was available for comparison and because the NEM encompasses such a
large fraction of Australia's electricity consumption.

Table 3
Description of selected studies.

Region Main Author and Year Number of Scenario Years Models/Tools used Dispatch model time-step Simulated duration
Scenarios

United States Jacobson et al., 2015 [51] 1 2050–2055 LOADMATCH with GATOR-GCMOM 30 s 6 years
Germany SRU, 2011 [34] 4 2050 REMix 1h 1 year
France Krakowski et al., 2016 4 2010–2050 MARKAL-TIMES Representative periods 40 years
[32]
Australia Elliston et al., 2012 [18] 1 2010 Python program 1h 1 year
Australia AEMO, [9] 2 2030 Unnamed Monte Carlo model and dispatch 1h 1 year
model
Portugal Krajačić et al., 2010 [43] 1 2020 H2RES 1h 1 year
Portugal Fernandes et al., 1 2020 EnergyPlan 1h 1 year
2013 [44]
New Zealand Mason et al., 2013 [42] 1 2005–2010 Matlab script with Excel 30 min 6 years

4
 W. Deason

Table 4
Comparison of scenario demand assumptions.

Scenario Scenario Demand Scenario Years % of 2013 Electricity % of 2013 Total Energy Scenario demand assumptions
(TWh)a Demandb Demandc

Jacobson et al. 2015 (USA) [51] 13782.5 2050–2055 366% 79% All transportation, industry, and heating (electric and hydrogen) with energy saving and
efficiency measures
SRU 2011 - Scenario 1.a (Germany) [34] 509 2050 98% 19% About half of auto fleet electrified and energy saving and efficiency measures
SRU 2011 - Scenario 1.b (Germany) [34] 700 2050 135% 27% Most of auto fleet electrified; large portion of heating and industry electrified with energy
saving and efficiency measures
SRU 2011 - Scenario 2.2.a (Germany) [34] 509.4 2050 98% 19% Same as Scenario 1.a
SRU 2011 - Scenario 2.2.b (Germany) [34] 700 2050 135% 27% Same as Scenario 1.b
Krakowski et al. 2016 - Scenario 100RES 2050 425.4 2010–2050 97% 23% Demand is considered elastic, but based on reference values which consider some
(France) [32] electrification, energy saving, and efficiency measures in all sectors
Krakowski et al. 2016 - Scenario 100RES 425.4 2010–2050 97% 23% Same as Scenario 100RES 2050
2050_v1 (France) [32]
Krakowski et al. 2016 - Scenario 100RES 425.4 2010–2050 97% 23% Same as Scenario 100RES 2050

5
2050_v5 (France) [32]
Krakowski et al. 2016 - Scenario 100RES 425.4 2010–2050 97% 23% Same as Scenario 100RES 2050
2050_v7 (France) [32]
Elliston et al. 2012 (Australia) [18] 204.4 2010 113% (NEM only)d 22% (All AUS) 2010 National Electricity Market (NEM) demand was used with no augmentation
AEMO 2013 - Scenario 1 in 2030 (Australia) 238 2030 131% (NEM only) 25% (All AUS) Moderate economic growth; rapid uptake of energy efficiency, six million EVs by 2030; NEM
[9] electricity use only
AEMO 2013 - Scenario 2 in 2030 (Australia) 270 2030 149% (NEM only) 29% (All AUS) High economic growth; moderate uptake of energy efficiency. 3.5 million EVs by 2030; NEM
[9] electricity use only
Krajačić et al. 2010 - Scenario 3 (Portugal) 49.176 2020 109% 26% 2006 Portuguese electricity load was used with no augmentation
[43]
Fernandes et al. 2013 - Scenario 4 (Portugal) 53.6 2020 118% 28% Forecasted demand based on principles of reduced energy dependence and sustainable
[44] electricity production
Mason et al. 2013 (New Zealand) [42] 41.1505 2005–2010 108% 27% 2005–2010 New Zealand electricity load was used with no augmentation

a
If the number of scenario years was greater than one, the average scenario demand/year was recorded.
b
Electricity consumption for 2013 was taken from [61], except for the Australian studies (NEM grid), for which 2013 electricity consumption was taken from [62].
c
Total energy consumption for countries in 2013 was taken from [60] and converted from Mtoe to TWh.
d
The NEM supplies around 80% of Australia's electricity consumption [59].
Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx
W. Deason Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx

Jacobson et al. 2015 (USA) Fig. 1. Comparison of the variable genera-

SRU 2011 - Scenario 1.a (Germany) tion capacity with the capacity for flexibility
SRU 2011 - Scenario 1.b (Germany) in the system. The data used to create this
SRU 2011 - Scenario 2.2.a (Germany) chart is exactly the same as the data used in
SRU 2011 - Scenario 2.2.b (Germany) Fig. 2. Capacity for flexibility is defined as
any generation, storage, or demand capacity
Krakowski et al. 2016 - Scenario 100RES 2050 (France)
that can temporally adjust to respond to
Krakowski et al. 2016 - Scenario 100RES 2050_v1 (France)
changes in output from variable generation
Krakowski et al. 2016 - Scenario 100RES 2050_v5 (France)
capacity or to changes in demand. Refer to
Krakowski et al. 2016 - Scenario 100RES 2050_v7 (France)
Table 4 for scenario reference numbers.
Elliston et al. 2012 (Australia)
AEMO 2013 - Scenario 1 in 2030 (Australia)
AEMO 2013 - Scenario 2 in 2030 (Australia)
Krajačić et al. 2010 - Scenario 3 (Portugal)
Fernandes et al. 2013 - Scenario 4 (Portugal)
Mason et al. 2013 (New Zealand)
0% 20% 40% 60% 80% 100%
Variable Generation Capacity
Fraction of Total Scenario Capacity
Capacity for Flexibility

capacity (SRU 2011 Scenarios 1.a and 1.b (Germany) feature 73% and generation, it could be used in a baseload fashion. This is particularly
69% respectively) [34]. This can be attributed to the high capacity apparent through the 2.2.a and 2.2.b scenarios in SRU 2011's study
factor for off-shore wind turbines, which generate a large amount of the when compared to the 1.a and 1.b scenarios [34]. Through the in-
electricity in the SRU scenarios [34]. troduction of the ability to import electricity, the need to use biofuels as
Fig. 1 also shows that for the examined scenarios, the fraction of a complement to variable generation could be reduced and biofuel
variable generation capacity increases with system size. This observa- generation capacity could be used in a more economical, baseload
tion is in alignment with other literature. Studies have shown that the fashion. The same strategy can be seen by comparing the hydro capa-
spacing between wind farms has been shown to effectively increase city factor. The capacity factor of hydro in the AEMO 2013 scenarios is
average capacity factors3 of wind generation [63,64] and aggregating reduced compared to the other scenarios due to its use in com-
electric load over large distances can effectively smooth demand [65]. plementing variable generation [9].
However, in this case the specific geographic factors and selected Figs. 4 and 5 show the different generation technologies used in the
technologies in the scenarios (discussed later in this section) may be an scenarios, along with reducible demand and transmission, providing
overriding factor. This is seen in the larger influence of hydroelectric further detail into the categories used in Fig. 2. Reducible demand is
generation in the Mason et al. and Portuguese scenarios [43,44], the shown alongside flexible generation and transmission capacities in
selection of concentrated solar thermal in the Australian scenarios Fig. 5 due to its ability to be implemented in place of those technolo-
[9,18], and the use of electricity import in the Krakowski et al. [32] and gies. Rather than call for an increase in output from a flexible genera-
SRU 2011 [34] scenarios, which are all grouped as capacity for flex- tion source or neighbouring system, a grid operator could call for a
ibility in Fig. 1 (further discussion of these technology makeups will large energy user to temporarily reduce output.
come later in the article). If these scenarios had similar insolation and Fig. 6 shows the different storage technologies, along with flexible
wind speed characteristics and had chosen similar system capacity demand, used in the scenarios. Flexible demand has been shown next to
makeups, these due to larger grids may have been discernible from the storage technologies because flexible demand can be seen as storing the
data. demand for later. In both cases, the total energy use essentially remains
A comparison of key capacities and how they compare with each the same (ignoring inefficiencies due to energy storage or variable en-
other and with historical generating capacity is shown in Fig. 2. The ergy usage rates), but the time at which it is used changes.
United States energy system as proposed by Jacobson et al. features a The large amount of hydrogen electrolysis capacity in the Jacobson
large amount of Total Scenario Capacity relative to the present day et al. scenario shown in Fig. 6 behaves in a similar way as the other
generation capacity (U.S. electricity generation system in 2013) due to flexible demand strategies [51]. When available, electricity is used to
the large demand of energy that must be met directly through elec- generate hydrogen; however, the capacity for electrolysis has to be
tricity or indirectly through hydrogen [51]. The other scenarios all large enough to make up for times when it is not. The inclusion of
show an increase in generation capacity as well. While increase in flexible and reducible demand within total system capacity is a unique
electricity demand can account for some of the capacity increase seen in addition in the present work. The inclusion of flexible and reducible
the scenarios, the majority of increase is due to decreased capacity demand is crucial in several scenarios for providing the flexibility re-
factors of the generation technology used compared with present day. quired to meet demand at all times.
Fig. 3 shows the relative capacity factors different technologies used No single technology is fully dominant for any scenario; rather, all
in the scenarios. Only some of the scenarios are included in the figure require a diverse portfolio of technologies working in sync to meet
due to lack of available information. For variable generation technol- demand. This is particularly apparent in the choice of flexible tech-
ogies, this is due to differences in the availability of energy for the nologies chosen by scenarios. By examining the technologies shown in
studied region. However, some of the flexible generation technologies Figs. 4–6 as a whole, several regional patters can be seen:
show a wide variation in their capacity factors. This can be attributed to
the importance the specific technology played in a particular scenario. 1. All the scenarios include wind technology. One German scenario,
For example, if it served as a direct complement for variable generation one French scenario, and the New Zealand scenario stand out by
(i.e. only operating when resources are not available for variable gen- exclusively using wind technology, with no PV capacity installed.
erators) then its capacity factor will be relatively low. However, if a 2. All the scenarios include hydroelectric and/or pumped storage
technology's capacity was not needed to complement variable technologies. The Portuguese and New Zealand scenarios stand out
in using it as a significant portion of their flexible capacity.
3. All the scenarios except for the American scenario include biomass
3
In this context, capacity factor is defined as average ratio of the actual output of a technology. The German, French, and Australian scenarios stand out
generation technology to its potential output. This is the definition used throughout this
as using it as a significant portion of their flexible capacity.
report.

6
W. Deason Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx

Jacobson et al. 2015 (USA) Fig. 2. Makeup of the total scenario capa-
SRU 2011 - Scenario 1.a (Germany) city for each scenario. The circle on each
SRU 2011 - Scenario 1.b (Germany) scenario bar represents the electric genera-
SRU 2011 - Scenario 2.2.a (Germany) tion capacity for the target country in the
SRU 2011 - Scenario 2.2.b (Germany) present day. Capacity information for the
Krakowski et al. 2016 - Scenario 100RES 2050 (France) Australian studies (NEM grid) comes from
Krakowski et al. 2016 - Scenario 100RES 2050_v1 (France) [66] and [67] and is for the year 2016. All
Krakowski et al. 2016 - Scenario 100RES 2050_v5 (France) other capacity information comes from [61]
Krakowski et al. 2016 - Scenario 100RES 2050_v7 (France) and is for the year 2013, except for Ger-
Elliston et al. 2012 (Australia) many, for which information on coal, oil,
AEMO 2013 - Scenario 1 in 2030 (Australia) natural gas, biofuels, and waste incineration
AEMO 2013 - Scenario 2 in 2030 (Australia) components of present day capacity is for
Krajačić et al. 2010 - Scenario 3 (Portugal) the year 2000, and for France, for which
Fernandes et al. 2013 - Scenario 4 (Portugal) biofuel and waste incineration capacity is
Mason et al. 2013 (New Zealand) not available, but assumed to be 1 GW
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% combined. Refer to Table 4 for scenario re-
Fraction of Total Scenario Capacity ference numbers.

Variable Generation Capacity Flexible Generation Capacity

Transmission Capacity Storage Capacity
Flexible Demand Reducible Demand
Present Day Generation Capacity

100% Fig. 3. Technology-specific capacity factors

for some of the scenarios. Refer to Table 4
Jacobson et al. 2015 (USA)
for scenario reference numbers.
SRU 2011 - Scenario 1.a (Germany)
75%
Capacity Factor (%)

SRU 2011 - Scenario 1.b (Germany)

SRU 2011 - Scenario 2.2.a (Germany)
50% SRU 2011 - Scenario 2.2.b (Germany)
AEMO 2013 - Scenario 1 in 2030 (Australia)
AEMO 2013 - Scenario 2 in 2030 (Australia)
25% Mason et al. 2013 (New Zealand)

0%

% of Total Fig. 4. Technology specific makeup of the

Jacobson et al. 2015 (USA) 57.31% variable generation capacity for each sce-
SRU 2011 - Scenario 1.a (Germany) 73.05% nario, including percentage of total scenario
SRU 2011 - Scenario 1.b (Germany) capacity. Refer to Table 4 for scenario re-
69.32%
ference numbers.
SRU 2011 - Scenario 2.2.a (Germany) 43.80%
SRU 2011 - Scenario 2.2.b (Germany) 57.29%
Krakowski et al. 2016 - Scenario 100RES 2050 (France) 52.45%
Krakowski et al. 2016 - Scenario 100RES 2050_v1 (France) 60.28%
Krakowski et al. 2016 - Scenario 100RES 2050_v5 (France) 50.62%
Krakowski et al. 2016 - Scenario 100RES 2050_v7 (France) 17.54%
Elliston et al. 2012 (Australia) 43.86%
AEMO 2013 - Scenario 1 in 2030 (Australia) 42.55%
AEMO 2013 - Scenario 2 in 2030 (Australia) 54.46%
Krajačić et al. 2010 - Scenario 3 (Portugal) 39.88%
Fernandes et al. 2013 - Scenario 4 (Portugal) 45.60%
Mason et al. 2013 (New Zealand) 22.97%
0% 20% 40% 60% 80% 100%
Fraction of Total Variable Generation Capacity

Solar PV Off Shore Wind On Shore Wind

Undefined Wind Ocean

4. The European scenarios on Germany, France, and Portugal all take production were only used in American scenarios
advantage of their regional location to import or export electricity b. Compressed air energy storage (CAES) was only used in the
through transmission lines. German scenarios.
5. There was no single storage technology used an every scenario, c. Hydrogen electricity generation and battery capacity were only
though, pumped hydro was used the most. used in the Portuguese scenarios.
6. Several flexible generation and storage technologies were not shared
between studies:
a. Solar heating, non-CSP solar storage, and hydrogen heat

7
W. Deason Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx

% of Total Fig. 5. Technology specific makeup of the

Jacobson et al. 2015 (USA) 9.86% flexible generation and transmission capa-
SRU 2011 - Scenario 1.a (Germany) 14.41% city and reducible demand for each sce-
SRU 2011 - Scenario 1.b (Germany) 18.80% nario, including percentage of total scenario
SRU 2011 - Scenario 2.2.a (Germany) capacity. Refer to Table 4 for scenario re-
47.60%
ference numbers.
SRU 2011 - Scenario 2.2.b (Germany) 36.52%
Krakowski et al. 2016 - Scenario 100RES 2050 (France) 34.80%
Krakowski et al. 2016 - Scenario 100RES 2050_v1 (France) 24.36%
Krakowski et al. 2016 - Scenario 100RES 2050_v5 (France) 37.10%
Krakowski et al. 2016 - Scenario 100RES 2050_v7 (France) 62.69%
Elliston et al. 2012 (Australia) 53.59%
AEMO 2013 - Scenario 1 in 2030 (Australia) 46.71%
AEMO 2013 - Scenario 2 in 2030 (Australia) 40.54%
Krajačić et al. 2010 - Scenario 3 (Portugal) 27.04%
Fernandes et al. 2013 - Scenario 4 (Portugal) 36.18%
Mason et al. 2013 (New Zealand) 62.75%
0% 20% 40% 60% 80% 100%

Fraction of Total Flexible Generation and Transmission Capacity and Reducible Demand
CSP Solar Heating Hydro
Geothermal Biofuels Transmission
Reducible Demand

% of Total Fig. 6. Technology specific makeup of the

Jacobson et al. 2015 (USA) 32.82% storage and flexible demand for each sce-
SRU 2011 - Scenario 1.a (Germany) 12.54% nario, including percentage of total scenario
SRU 2011 - Scenario 1.b (Germany) 11.88% capacity. Refer to Table 4 for scenario re-
SRU 2011 - Scenario 2.2.a (Germany) ference numbers.
8.60%
SRU 2011 - Scenario 2.2.b (Germany) 6.19%
Krakowski et al. 2016 - Scenario 100RES 2050 (France) 12.75%
Krakowski et al. 2016 - Scenario 100RES 2050_v1 (France) 15.36%
Krakowski et al. 2016 - Scenario 100RES 2050_v5 (France) 12.28%
Krakowski et al. 2016 - Scenario 100RES 2050_v7 (France) 19.77%
Elliston et al. 2012 (Australia) 2.55%
AEMO 2013 - Scenario 1 in 2030 (Australia) 10.74%
AEMO 2013 - Scenario 2 in 2030 (Australia) 5.00%
Krajačić et al. 2010 - Scenario 3 (Portugal) 33.08%
Fernandes et al. 2013 - Scenario 4 (Portugal) 18.21%
Mason et al. 2013 (New Zealand) 14.27%
0% 20% 40% 60% 80% 100%

Fraction of Total Storage Capacity and Flexible Demand

Non-CSP Thermal Storage Hydrogen Heat Production Capacity
Compressed Air Energy Storage (CAES) Pumped Storage
Hydrogen Electricity Generation Capacity Battery Capacity
Undefined Storage Flexible Demand: Demand Side Participation
Flexible Demand: EV Flexible Demand: Hydrogen Electrolysis

5.3. Scenario cost Krakowski et al. scenarios appears to be due to the cost of providing
backup electricity from either electricity import or through construc-
Table 5 shows the projected scenario electricity costs provided by tion of biomass plants, which saw reduced load hours [32]. Relative to
some of the studies. For the studies by Jacobson et al. [51], Krakowski the baselined Krakowski et al. scenario, v1 removed the restriction of a
et al. [32], and AEMO [9], costs for transmission (distribution and low wind output period in January; v5 lessened the restriction on the
delivery) were included in the number provided. These values can be potential for biomass availability, while v7 set upper limits on the
compared to the generation makeup shown in Figs. 4, 5, and 6 to gain amount of variable generation allowed. The costs reflect the need for
perspective of the contribution of cost to each scenario. Additionally, 4 more or less amounts of complementary flexible generation capaci-
out of the 5 studies also provided baseline or business-as-usual (BAU) ty—thus, a higher or lower cost—relative to the reference scenario.
electrical generation costs to compare their scenario costs against. By comparing the scenarios with their respective BAU electricity
These numbers are shown in Table 5 alongside their respective scenario costs, it can be seen that 3 out of 4 studies predict between 41% and
electricity costs. The 2012 electricity price for industry has been pro- 104% higher electricity costs in the 100% RE scenarios. The Jacobson
vided for comparison as well. et al. study differs from the others in predicting that costs are the same
Details within the studies enable some comparison of the costs in both the BAU and 100% RE scenarios, and 62% cheaper when ex-
within the scenarios of individual studies. The primary reason for in- ternalities are included [51]. No pattern could be seen between the cost
creased costs for two of the four SRU scenarios appears to be the use of values give in the study and the reported 2012 electricity prices for the
large amounts of solar PV and offshore wind for providing electricity consumer, including the BAU cost values. This mismatch hints that
generation [34]. According to the study, these technologies are more there may be other variables not being taken into account by the studies
expensive to construct than importing electricity, as was done in the in their cost calculations.
cheaper scenarios. The reason for increased costs between the

8
 W. Deason

Table 5
Projected generation costs for some of the scenarios.

Scenario Scenario Electricity Costsa Scenario Electricity Costsa BAU Electricity Costs % Increase relative to BAU Electricity Current Day Electricity Price for Industry in Scenario
Costs Countryb
2012USD /Year 2012USD /MWh 2012USD /MWh 2012USD/MWh

Jacobson et al. 2015 (USA) [51] 1.54 Trillion 104c 104 (271)d 0% (−62%) 66.8 (Incl. tax)
SRU 2011 - Scenario 1.a (Germany) [34[ 62.6 Billion 123 N/A N/A 148.7 (Ex-tax)
SRU 2011 - Scenario 1.b (Germany) [34] 110 Billion 157 N/A N/A 148.7 (Ex-tax)
SRU 2011 - Scenario 2.2.a (Germany) [34] 45.2 Billion 89 N/A N/A 148.7 (Ex-tax)
SRU 2011 - Scenario 2.2.b (Germany) [34] 69.0 Billion 98 N/A N/A 148.7 (Ex-tax)
Krakowski et al. 2016 - Scenario 100RES 2050 (France) 80.4 Billion 189 110 72% 121.7 (Ex-tax)
[32]
Krakowski et al. 2016 - Scenario 100RES 2050_v1 65.9 Billion 155 110 41% 121.7 (Ex-tax)
(France) [32]

9
Krakowski et al. 2016 - Scenario 100RES 2050_v5 68.5 Billion 161 110 46% 121.7 (Ex-tax)
(France) [32]
Krakowski et al. 2016 - Scenario 100RES 2050_v7 81.7 Billion 192 110 75% 121.7 (Ex-tax)
(France) [32]
AEMO 2013 - Scenario 1 in 2030 (Australia) (NEM Only) 27.4 Billion 115 57e 102% 172.3 (Ex-tax) (All Australia)
[9]
AEMO 2013 - Scenario 2 in 2030 (Australia) (NEM Only) 31.4 Billion 116 57 104% 172.3 (Ex-tax) (All Australia)
[9]
Fernandes et al. 2013 - Scenario 4 (Portugal) [44] 4.84 Billion 90.3 45 101% 147.3 (Ex-tax)

a
Costs have been converted to 2012USD from their local currency and study year [69,70].
b
The 2012 electricity price for industry was taken from [68] and was chosen for comparison due to its assumed proximity to wholesale electricity price. If available, the ex-tax price was chosen for the same reason.
c
Values in bold were available in their current form (cost/MWh or cost/year) in the original publication, while unbolded cost values were derived from the other form (cost/MWh or cost/year) using the total energy generation in the scenario
year.
d
The value in parentheses denotes a BAU cost including externalities (health and climate costs).
e
Provided in AEMO as an estimate for the 2012 wholesale electricity cost [9].
Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx
W. Deason Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx

6. Conclusions References

The compared scenarios showed that many different strategies and [1] Welsch M. Enhancing the treatment of systems integration in long-term energy
technologies can be used to provide flexibility in a 100% RE system. To models (PhD thesis). Stockholm, Sweden: KTH Royal Institute of Technology; 2013.
[2] German Energy Agency. dena ancillary services study. Berlin, Germany: dena; 2014.
complement the variability of electricity generation from variable [3] Loftus PJ, Cohen AM, Long JCS, Jenkins JD. A critical review of global dec-
generation sources, energy must be generated from a flexible genera- arbonization scenarios: what do they tell us about feasibility? WIREs Clim Change
tion source, stored in some fashion, or if possible, demand must be 2015;6(1):93–112.
[4] WWF International, Ecofys, Office for Metropolitan Architecture. The energy report.
adjusted. To facilitate comparison of scenarios, the primary energy Gland, Switzerland: WWF International; 2011.
production from renewables was grouped in the category ‘Variable [5] Jacobson M, Delucchi M. Providing all global energy with wind, water, and solar
Generation Capacity’, while all other mechanisms supporting flexibility power, part I; technologies energy resources, quantities and area of infrastructure,
and materials. Energy Policy 2011;39:1154–69.
(generation, storage, flexible demand, etc.) were grouped in the cate- [6] Delucchi M, Jacobson M. Evaluating the feasibility of meeting all global energy
gory ‘Capacity for Flexibility’. Together, these two categories compose needs with wind, water, and solar power, part II: reliability, system and transmis-
the principle category, ‘Total Scenario Capacity’. By including capa- sion costs, and policies. Energy Policy 2011;39:1170–90.
[7] Sawin JL, Moomaw WR. Renewable revolution: low-carbon energy by 2030.
cities for storage and flexible demand in the total scenario capacity
Washington, D.C., USA: Worldwatch Institute; 2009.
category, the 100% renewable energy system scenarios can be better [8] Reedman LJ. High penetration renewables studies: a review of the literature, Report
compared with each other and with present day generation capacity. No. EP 127113. Canberra, Australia: Commonwealth Scientific and Industrial
Flexibility was found to be a crucial and significant component of a Research Organisation; 2012.
[9] AEMO. 100 per cent renewables study - modelling outcomes. Australia: Australian
100% RE system. None of the scenarios reviewed in the present work Energy Market Operator (AEMO); 2013.
were able to provide all of their energy using only variable generation [10] National Renewable Energy Laboratory. Renewable energy futures study, NREL/TP-
capacity, with only two of the scenarios capable of reaching over 65% 6A20-52409. Golden, Colorado: NREL; 2012.
[11] Seligman P. Australian sustainable energy - by the numbers. Melbourne, Australia:
variable generation penetration of total scenario capacity (SRU 2011 Melbourne Energy Institute, University of Melbourne; 2012.
Scenarios 1.a and 1.b (Germany) [34] feature 73% and 69% respec- [12] Wright M, Hearps P. Australian sustainable energy - zero carbon Australia sta-
tively). tionary energy plan. Melbourne, Australia: Beyond Zero Emissions; 2010.
[13] Cochran J, Mai T, Bazilian M. Meta-analysis of high penetration renewable energy
Predicted electricity costs from the scenarios varied drastically in all scenarios. Renew Sustain Energy Rev 2014;29:246–53.
regions, and from these results, it was not possible to conclusively say [14] Lund H, Mathiesen B. Energy system analysis of 100% renewable energy systems:
that a 100% RE system from one region is less economic than another. the case of Denmark in years 2030 and 2050. Energy 2009;34:524–31.
[15] Greenpeace International. European renewable energy council, global wind energy
Each calculated cost is highly dependent upon the specific environment council, 4th Edition 2012 World Energy Scenario. Amsterdam, The Netherlands:
of each scenario and the cost calculation method used. Lack of in- Greenpeace International; 2012.
formation on the specific cost components and methodologies used (e.g. [16] Schmid E, Pahle M, Knopf B. Renewable electricity generation in Germany: a meta-
analysis of mitigation scenarios. Energy Policy 2013;61:1151–63.
unit costs for capacity, unit lifetimes, generation statistics, and cost
[17] Trainer T. Can Australia run on renewable energy? The negative case. Energy Policy
calculation methods) prevented direct comparison. However, some 2012;no. 50:306–14.
studies provided multiple scenarios (including business-as-usual sce- [18] Elliston B, Diesendorf M, MacGill I. Simulations of scenarios with 100% renewable
narios) that could be compared. In 3 out of the 4 studies including a electricity in the Australian National Electricity Market. Energy Policy
2012;45:606–13.
business-as-usual cost, the 100% RE scenarios were found to be be- [19] Huva R, Dargaville R, Rayner P. Optimising the deployment of renewable resources
tween 41% and 104% more expensive. The fourth study (Jacobson et al. for the Australian NEM (National Electricity Market) and the effect of atmospheric
[51]) showed electricity costs to be the same for a 100% RE system and length scales. Energy 2016;no. 96:468–73.
[20] Schmidt J, Cancella R, Junior A Olimpio Pereira. An optimal mix of solar PV, wind
business-as-usual system, and the 100% RE system to be 62% cheaper and hydro power for a low-carbon electricity supply in Brazil. Austria: Universtaet
when externalities are included. fuer Bodenkultur Wien, Vienna; 2015.
The scenarios discussed in the present work and the strategies used [21] Krajačić G, Duić N, da Graҫa Carvalho M. H2RES, energy planning tool for island
energy systems - The case for the island of Mljet. Int J Hydrog Energy
in those scenarios are not exhaustive of all scenarios possible in a 2009;34:7015–26.
decarbonized future. Though not examined in this study, further in- [22] Lund H, Mathiesen B. Energy system analysis of 100% renewable energy
tegration of different energy carriers, usage or markets (e.g. heat and systems—The case of Denmark in years 2030 and 2050. Energy 2009;34:524–31.
[23] Mathiesen BV, Lund H, Connolly D, Wenzel H, Østergaard PA, Möller B, et al. Smart
electricity produced from renewable, nuclear or concentrating solar
Energy systems for coherent 100% renewable energy and transport solutions. Appl
thermal sources, transportation, water desalination plants, etc.) will be Energy 2015;145:139–54.
needed to achieve full decarbonization. Some renewable technologies [24] Østergaard PA, Andersen FM, Kwon PS. Energy systems scenario modelling and
long term Forecasting of hourly electricity demand. Int J Sustain Energy Plan
and scenarios discussed in this study may not come to fruition because
Manag 2015;7:99–116.
of economic barriers or environmental changes, notably due to climatic [25] EREC. Re-thinking 2050 - a 100% renewable energy vision for the European Union.
changes. Brussels, Beligum: European Renewable Energy Council; 2010.
Future work could investigate the role and economic viability of [26] Greenpeace. Energy [r]evolution - towards a fully renewable energy supply in the
EU 27. Amsterdam,: Greenpeace International and European Renewable Energy
other low-carbon technologies (nuclear power or fossil with carbon Council; 2010.
capture and storage) to operate in a high-penetration RE system as a [27] Trainer T. Can Europe run on renewable energy? A negative case. Energy Policy
flexible generation source. Focus could also be given on the increased 2013;no. 63:845–50.
[28] Bertsch J, Growitsch C, Lorenczik S, Nagl S. Flexibility in Europe's power sector - An
flexibility brought to the energy system through the capability for these additional requirement or an automatic complement? Energy Econ
technologies to function as cogeneration systems (providing both useful 2016;53:118–31.
heat and electricity on-demand). [29] Sørenson B. A renewable energy and hydrogen scenario for northern Europe. Int J
Energy Res 2007;32(5):471–500.
[30] Dominković DF, Bačeković I, Ćosić B, Krajačić G, Pukšec T, Duić N, et al. Zero
carbon energy system of South East Europe in 2050. Appl Energy
Acknowledgements 2016;184:1517–28.
[31] Tröster E, Kuwahata R, Ackermann. European grid study 2030/2050. Langen,
Germany: Nergynautics GmbH; 2011.
I would like to acknowledge the United States Department of State [32] Krakowski V, Assoumou E, Mazauric V, Maizi N. Feasible path toward 40–100%
for supporting my Junior Professional Officer position at the renewable energy shares for power supply in France by 2050: a prospective ana-
lysis. Appl Energy 2016;171:501–22.
International Atomic Energy Agency. I would also like to acknowledge
[33] Drouineau M, Assoumou E, Mazauric V, Maïzi N. Increasing shares of intermittent
my colleagues, Bertrand Magné, Manuel Welsch, Aliki van Heek, and sources in Reunion Island: impacts on the future reliability of power supply. Renew
David Shropshire of the Planning and Economic Studies Section in their Sustain Energy Rev 2015;46:120–8.
guidance in the development of this project and helpful assistance with [34] SRU. Pathways towards a 100% renewable energy system. Berlin: German Advisory
Council on the Environment (SRU); 2011.
reviewing and improving it.

10
W. Deason Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx

[35] DLR. Long-term scenarios and strategies for the deployment of renewable energies 2009;37:387–99.
in Germany in view of European and global developments - Summary of the final [54] Budischak C, Sewell D, Thomson H, Mach L, Veron D, Kempton W. Cost-minimized
report. Stuttgart, Germany: German Aerospace Center; 2012. combinations of wind power, solar power and electrochemical storage, powering
[36] Allen P. Zero carbon Britain: rethinking the future. Machynlleth, Wales, United the grid up to 99.9% of the time. J Power Sources 2013;225:60–74.
Kingdom: Centre for Alternative Technology; 2013. [55] Short W, Diakov V. Matching Western US electricity consumption with wind and
[37] WWF-India and WISE. The energy report - Kerala - 100% renewable energy by solar resources. Wind Energy 2013;16:491–500.
2050. New Delhi, India: World Wildlife Foundation - India and World Institute for [56] Trainer T. Can the world run on renewable energy? A revised negative case.
Sustainable Energy; 2013. Humanomics 2013;29(2):88–104.
[38] Connolly D, Lund H, Mathiesen BV, Leahy M. The first step towards a 100% re- [57] Teske S, Sawyer S, Schäfer O. Energy [r]evolution - a sustainable world energy
newable energy-system for Ireland. Appl Energy 2011;88:502–7. outlook 2015. Amsterdam, The Netherlands: Global Wind Energy Council,
[39] Lehmann H. Energy rich Japan. Tokyo, Japan: Institute for Sustainable Solutions SolarPowerEurope; 2015.
and Innovations (ISUSI); 2003. [58] Jacobson MZ, Delucchi MA, Bauer ZAF, Goodman SC, Chapman WE, Cameron MA,
[40] Esteban M, Zhang Q, Utama A. Estimation of the energy storage requirement of a et al. 100% clean and renewable wind, water, and sunlight (WWS) all-sector energy
future 100% renewable energy system in Japan. Energy Policy 2012;47:22–31. roadmaps for 139 countries of the world. Stanford, California, United States:
[41] Ćosić B, Krajačić G, Duić N. A 100% renewable energy system in the year 2050: the Stanford University; 2016.
case of Macedonia. Energy 2012;48:80–7. [59] Australian Energy Market Commission. AEMC - National Electricity Market
[42] Mason IG, Page SC, Williamson AG. Security of supply, energy spillage control and [Online]. Available: 〈http://www.aemc.gov.au/Australias-Energy-Market/
peaking options within a 100% renewable electricity system for New Zealand. Markets-Overview/National-electricity-market〉.
Energy Policy 2013;60:324–33. [60] IEA. Energy balances of OECD countries: 2015 edition. Paris: International Energy
[43] Krajačić G, Duić N, da Graҫa Carvalho M. How to achieve a 100% RES electricity Agency; 2015.
supply for Portugal? Appl Energy 2011;88:508–17. [61] IEA. Electricity information 2015. Paris, France: International Energy Agency;
[44] Fernandes L, Ferreira P. Renewable energy scenarios in the portuguese electricity 2015.
system. Guimaraes: University of Minho; 2013. [62] AEMO. National electricity forecasting report for the national electricity market.
[45] Teske S. The advanced energy [r]evolution - a sustainable energy outlook for South Australian Energy Market Operator; 2014.
Africa. Amsterdam, The Netherlands: Greenpeace and European Renewable Energy [63] Alexander M, James P, Richardson N. Energy storage against interconnection as a
Council; 2011. balancing mechanism for a 100% renewable UK electricity grid. IET Renew Power
[46] de Otto S, Bevacqua M. Energy [R]evolution for the Canary Islands - Energy design Gener 2014;9(2):131–41.
for sustainable islands. Madrid, Spain: Greenpeace; 2015. [64] Santos-Alamillos FJ, Pozo-Vazquez D, Ruiz-Arias JA, Lara-Fanego V, Tovar-
[47] Peter S, Doleschek A, Lehmann H, Mirales J, Puig J, Corominas J, et al. Solar cat- Pescador J. Analysis of spatiotemporal balancing between wind and solar energy
alonia - a pathway to a 100% renewable energy system for Catalonia. Fundació resources in the southern Iberian Peninsula. J Appl Meteorol Climatol
Terra, iSuSI, and Ecoserveis; 2007. 2012;51:2005–24.
[48] Gustavsson M, Särnholm E, Stigson P, Zetterberg L. The IVL scenario: energy sce- [65] Corcoran BA, Jenkins N, Jacobson MZ. Effects of aggregating electric load in the
nario for Sweden 2050 - based on renewable energy technologies and sources. United States. Energy Policy 2012;46:399–416.
Göteborg, Sweden: IVL Swedish Environmental Research Institute Ltd.; 2011. [66] Australian Energy Regulator, Generation capacity and output by fuel source
[49] Pfenninger S, Keirstead J. Renewables, nuclear, or fossil fuels? Scenarios for Great [Online]. Available: 〈https://www.aer.gov.au/wholesale-markets/wholesale-
Britain's power system considering costs, emissions and energy security. Appl statistics/generation-capacity-and-output-by-fuel-source〉. [Accessed 01 December
Energy 2015;152:83–93. 2016].
[50] Mai T, Wiser R, Sandor D, Brinkman G, Heath G, Denholm P, et al. "Volume 1: [67] Australian Energy Regulator, Generation capacity and peak demand [Online].
exploration of high-penetration renewable electricity futures. Golden, Colorado: Available: 〈https://www.aer.gov.au/wholesale-markets/wholesale-statistics/
National Renewable Energy Laboratory; 2012. generation-capacity-and-peak-demand〉. [Accessed 01 December 2016].
[51] Jacobson MZ, Delucchi MA, Cameron MA, Frew BA. Low-cost solution to the grid [68] International Energy Agency. Energy prices and taxes: quarterly statistics - first
reliability problem with 100% penetration of intermittent wind, water, and solar for quarter 2013. Paris, France: IEA; 2013.
all purposes. PNAS 2015;112(49):15060–5. [69] The World Bank. Inflation, consumer prices (annual %), World Development
[52] Jacobson MZ, Delucchi MA, Bazouin G, Bauer ZAF, Heavey CC, Fisher E, et al. 100% Indicators, World DataBank, 〈https://data.worldbank.org/indicator/FP.CPI.
clean and renewable wind, water, and sulight (WWS) all-sector energy roadmaps TOTL.ZG〉.
for the 50 United States. Energy Environ Sci 2015;8:2093–117. [70] US Treasury Department. Treasury reporting rates of exchange as of December 31,
[53] Fthenakis V, Mason JE, Zweibel K. The technical, geographical, and economic 2012; 29 April 2015. [Online]. Available: 〈https://www.fiscal.treasury.gov/
feasibility for solar energy to supply the energy needs of the US. Energy Policy fsreports/rpt/treasRptRateExch/1212.pdf〉.

11