Fossil fuel depletion and socio-economic scenarios: An integratedapproach
I~nigo Capell�an-P�erez a, *, Margarita Mediavilla b, Carlos de Castro c, �Oscar Carpintero d,Luis Javier Miguel b
a Low Carbon Programme, Instituto de Economía Pública, University of Basque Country, Avd. Lehendakari Aguirre, 48015 Bilbao, Spainb Systems Engineering and Automatic Control, Escuela de Ingenierías Industriales, Paseo del Cauce s/n, University of Valladolid, 47011 Valladolid, Spainc Applied Physics Department, Escuela de Arquitectura, Av Salamanca, 18, University of Valladolid, 47014 Valladolid, Spaind Applied Economics Department, Facultad de Ciencias Econ�omicas, Paseo del Cauce, s/n, University of Valladolid, 47011 Valladolid, Spain
a r t i c l e i n f o
Article history:Received 14 January 2014Received in revised form12 September 2014Accepted 14 September 2014Available online 25 October 2014
The progressive reduction of high-quality-easy-to-extract energy is a widely recognized and alreadyongoing process. Although depletion studies for individual fuels are relatively abundant, few of themoffer a global perspective of all energy sources and their potential future developments, and even fewerinclude the demand of the socio-economic system.
This paper presents an Economy-Energy-Environment model based on System Dynamics which in-tegrates all those aspects: the physical restrictions (with peak estimations for oil, gas, coal and uranium),the techno-sustainable potential of renewable energy estimated by a novel top-down methodology, thesocio-economic energy demands, the development of alternative technologies and the net CO2
emissions.We confront our model with the basic assumptions of previous Global Environmental Assessment
(GEA) studies. The results show that demand-driven evolution, as performed in the past, might be un-feasible: strong energy-supply scarcity is found in the next two decades, especially in the transportationsector before 2020. Electricity generation is unable to fulfill its demand in 2025e2040, and a largeexpansion of electric renewable energies move us close to their limits. In order to find achievable sce-narios, we are obliged to set hypotheses which are hardly used in GEA scenarios, such as zero or negativeeconomic growth.
In recent years, concerns about the depletion of energy andmaterials (e.g. peak oil), as well as limits to the ecosystem'sassimilation capacity of residues (e.g. climatic change) have beenraised in the social, political and business arena. All fossil fuels andminerals are finite and non-renewable on a human scale. Theseresources are thus limited physically and, more stringently,economically. However, different views about this phenomenonexist in the scientific discussion, opposing “geologists” (or pessi-mists) vs. “conventional economists” (or optimists). The first [67]argue that geological factors determine a peak in the extraction
of each resource that technology can only slightly modify e see forexample [22,81,128] and the activity of ASPO in http://www.peakoil.net and point out that these restrictions might havestrong economic consequences [19,56,61,99,136]. However, the“conventional economists”, applying the basis of neoclassicalgrowth theory [127], claim that market mechanisms and humaningenuity will be able to both transform resources into reserves andfind alternative energy sources to replace the scarce ones at asufficient pace to avoid supply restrictions, and thus, not affect GDPgrowth [1,87,104,122,132]. This paper intends to shed light on thisdiscussion by using a System Dynamics (SD) model that includesboth the physical data of the energy resources and the economicdata.
The fact that the peak of conventional oil has already occurredhas been largely admitted in Academia (e.g. Ref. [100]) as well as by
2 For example, some international economic organizations such the OECD [105]project that global GDP will grow at around 3% per year over the next half cen-tury. It will almost triple in the years 2010e2060, although world GDP distribution
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666642
international institutional agencies -together with the acknowl-edgment of peak oil basic theory as an appropriate methodology e
Refs. [13,147,148,150], representing government declarations (e.g.the European Energy Commissioner1 in 2009) and even from someoil companies [97]. In 2012, the ratio of oil in the global energyconsumption mix fell to its minimumvalue in the last 50 years [16].Annual oil discoveries peaked in the 60s and no oil price rise sincethen could invert or stop the tendency of declining discoveriesthereafter. Due to the close relation between natural gas and oil, thegeological understanding of their deposits and depletion is verysimilar. Conventional wisdom has it that global coal and uraniumreserves are ample and supply restrictions due to scarcity must notbe expected within the next several decades or even this century,but this is disputed by several studies [35,40,60,96,115].
Consequently, renewable energy, and particularly solar andwind energy, are the two main sources of renewable energy whichmight substitute the decline in fossil fuel extraction [124]. However,recent studies of their limits show that their potentials might beeven lower than the current final consumption of energy by meansof fossil fuels [30e32]. Thus, if a long-term structural scarcity inenergy supplies in the next few years and/or decades occurs, assuggested in the past (e.g. Reports from the Club of Rome [90,91],Global 2000 [11]), and, more recently [29,85,102,142], this situationwould be unprecedented in modern history. Moreover, the study ofprevious technological transitions shows that they are slow, in theorder of decades [49].
On the other hand, energy consumption acts as a climaticchange driver [70]. But few studies have focused on the effect ofenergy constraints in climate scenarios, e.g. Refs. [18,64,145].
While depletion estimation for individual fuels followingdifferent approaches are relatively abundant (see Refs. [85,93] foran overview), few studies have centered on the objective of giving acomprehensive study, including estimates for all fossil fuels:Refs. [2,42,81,85,95,142] and even fewer have analyzed the wholesystem and fuel interactions [29,102,151], as we propose with ourmodel.
This paper intends to shed light on these issues by describingand showing the results of the model we have developed, WoLiM(World Limit Model), which is a continuation of previous SystemDynamic models developed [29,93]. WoLiM is a structurally-simpleand transparent model that compares data from many differentsources and helps to view global panoramas. The SD approach al-lows the combination of different kinds of variables from differentknowledge sources, such as socio-economic, geological and tech-nological, so they can be managed and integrated. The model in-cludes the exhaustion patterns of non-renewable resources andtheir replacement by alternative energies, the estimations of thedevelopment and market penetration of alternative technologies,the energy demand of the World's economy under different socio-economic scenarios, the sustainable potential of renewable en-ergies, and the estimations of CO2 emissions related to fossil fuelconsumption, all of them viewed in a dynamic framework.
On the other hand, scenario methodology offers an approach todeal with the complexity and uncertainty inherent to these inter-related issues and has become very popular in recent GlobalEnvironmental Assessments (GEAs), e.g. IPCC's Assessment reports[69,70,75], UNEP's Global Environmental Outlook [137,138,140] orthe Millennium Ecosystem Assessment (MEA) [89]. Each storylineentails the representation of a plausible and relevant story abouthow the future might unfold. We judge that this methodology is anadequate one for the design of the socio economic scenarios that
1 Original post deleted. A transcription can be found in <http://europe.theoildrum.com/node/5397>.
are needed as inputs to our model. The paper, therefore, quantifiesand implements five representative storylines identified in GEAstudies (as described in Ref. [144] and use them as input policies ofthe WoLiM model. By using this methodology, we replicate the usualvisions of the future explored by these international agencies, allowingthem to be confronted with the case of the energy development con-straints. In fact, to date, these international scientific bodies havelargely ignored these constraints [3,26,64].2
The paper is organized as follows: Section 2 overviews themodel and its main hypothesis and limitations. Section 3 describesthe modeling of non renewable and renewable resources. Section 4explains the estimation of energy demand and Section 5 describesthe calculation of CO2 emissions. Scenarios and results aredescribed in Sections 6 and 7. Finally, conclusions are drawn inSection 8.
2. Overview of the WoLiM model
In recent decades, many global energy-economy-environmentmodels, most focusing on climate change analysis, have beendeveloped (e.g. MESSAGE [101], IMAGE [15], MERGE [86], etc.),some based on systemdynamics [12,28,46]. However, most of thesemodels tend to use (very) large resource estimates [88] that aresubject to high uncertainties and are strongly biased towardsoverestimation due to the preeminence of optimistic economicassumptions [26,64,115]. Thus, few models explicitly recognizeresource limits such as peak oil and relate them to the economicgrowth [29,51,90,91,93]. The WoLiM model does recognize suchlimits and adopts the approach of URR (Ultimately RecoverableResources), which is an expert-estimate of the total amount ofresource that will ever be recovered and produced.
TheWoLiMmodel,3 which continues previous works [29,33,93],includes the following trends in a dynamic framework:
� The exhaustion patterns of non-renewable resources (URRapproach and maximum extraction curves).
� The replacement of non renewable by alternative energies.� The energy demand of the World's economy under differentsocio-economic scenarios.
� The sustainable potential of renewable energies.� The net CO2 concentrations.
WoLiM is based on a lineal structure (see Fig. 1) which starts bychoosing a scenario framework that consists of a set of socioeco-nomic and technological assumptions and policies that are inte-grated in a coherent and sensibleway (this scenariomethodology isdescribed in Section 6). The projection of the socio-economicdrivers establishes the world energy demand. This demand isthen disaggregated by sectors according to the different end-usesectors (electricity, industry, transport, etc.), and the energy de-mand of each sector is disaggregated into the demand by resources(liquid fuels, gas, electricity, etc.). These demands are compared tothe production of each particular resource, which is limited by thegeology-based peaks and the rates of technological substitution.Finally, the net CO2 emission and concentration levels arecomputed.
among countries will be very different from now: China and India will togetheraccount for 46 % of global GDP in 2050, up from less than 13 % today. No mention,however, about how scientific knowledge on resources constraints, likely futurescarcities and some other economic uncertainties may affect these forecasts.
3 For a full description of the WoLiM model see Ref. [24].
Fig. 1. Basic logic functioning of the WoLiM model.
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666 643
Although the model is based on System Dynamics (SD), thisversion is not a feedback-rich one as SD models tend to be. Some ofthe most obvious feedback loops are missing: e.g. it does notconsider that the energy scarcity would influence the economy, i.e.the lack of energy would, for example, reduce economic growth(e.g. Refs. [8,61]). This makes WoLiM, basically, a dynamic model ofenergy demand and technology trends versus physicalrestrictions.4
The reason for the simplification in this model version is the lackof consensus in the literature about the quantification of the impactof energy scarcity on the future economic growth. Although someauthors analyze this relationship (e.g. Ref. [61]), there is no well-developed and widely accepted theory on this topic. Despiteincreasing criticism on this dominant view [8,14], most macroeco-nomic models still pay very little attention to natural resources and,from our point of view, overestimate the capacity of technologicalsubstitution. Our studies are showing that the technological substi-tution, in particular the substitution of oil in this decade, is verydifficult and does not seem possible in a demand-driven transition[33,93]. On the other hand, the integration of this feedback tends todrive the system into collapse [33,102]. Therefore, prior to modelingan inadequate feedback, we decided not to include it.
Therefore, the relationship between economy and energy in ourmodel can be described as dual:
- Demand-driven if there is no restriction to the access of re-sources. In this case, the supply of energy is assumed to adjust tothe estimated demand.
- Supply-driven if the energy demand cannot be satisfied. In such acase, the estimated energy demand exceeds supply and an en-ergy scarcity would appear. Of course, in reality there would bean adjustment through a price increase to reach a new equi-librium, but the model cannot simulate it because that feedbackloop is missing, it only observes a discrepancy between demandand production.
Thus, the WoLiM model outputs are only valid while noimportant disequilibrium between demand and supply is reachedin any sector. Afterwards, the system could in fact evolve in a va-riety of ways and the results shown are unrealistic. However, itsmain contribution is the fact that it enables the detection of thoseyears and sectors (or “scarcity points”, as they are called in Section7) when the energy supply cannot meet the demand (under thegiven socioeconomic scenarios).
4 Another significant feedback, such as the impact of climate change on theeconomy, is not considered either, but there are some important loops includedwhich make WoLiM a feedback model, such as the fossil fuel extraction and therenewable energy dynamics. See Appendixes B and C for more detailedexplanations.
2.1. Key variables of the model
Exogenous variables are set by the scenario methodology, whileendogenous variables are calculated within the model. The exog-enous variables (or policies) are:
� GDP per capita growth.� Population growth.� Sectoral efficiency improvements (improvement of the energyintensity of the following economic sectors: transportation, in-dustry, electricity and buildings).
� Non-renewable extraction curves for oil, gas, coal and uranium.� Techno-sustainable potential of renewable energy sources.� Growth of renewable energies for electricity production (wind,solar PV and CSP, hydroelectric, geothermal, biomass & wasteand oceanic), and growth of nuclear power infrastructure.
� Growth of renewable energy for thermal uses and savingsrelated to efficiency in industry and buildings.
� Market penetration of alternative transport bymeans of electric,hybrid and gas vehicles.
� Market penetration of alternatives to liquid fuels by coal toliquids, gas to liquids and biofuels (first and second generation).
� Afforestation programs.
The number of endogenous variables of the model is large (over420), but the main ones are:
� Energy intensities of each economic sector: transportation,electricity, industry and buildings.
� Energy demands of each fuel (liquid fuels, gas, electricity, etc.)by sector. In order to find out the share of each fuel, historicaltrends have been extrapolated (unless a specific policy isapplied).
� Stocks and flows of non renewable resources (oil, gas, ura-nium, coal) whose depletion dynamics are described inAppendix B.
� Stocks that describe the infrastructure of renewable energies(solar, wind, hydroelectric, etc.) whose growth is determinedby the policies applied (see Appendix C for a detaileddescription).
� Stocks that represent the introduction of the alternative policies(biofuels, EV, efficiency, etc.), described in Appendix D.
� CO2 emissions and CO2 concentration levels related to fossil fueluse.
An overview of themodel (Forrester diagram) can be seen in Fig.2 and Appendix A, Fig. A1.
2.2. Assumptions and hypotheses
The main hypotheses of the model are the following:
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666644
� Non-renewable resources extraction is subject to geologicalconstraints (e.g. peak oil theory).
� Technological changes, such as the replacement of non renew-able by alternative energies or efficiency, take time. The transi-tion growth ratios are determined based on the tendenciesobserved in past decades (or accelerated under specific policies).
� The energy demand of the World's economy is determined bythe sectoral energy intensities, whose evolution is considered tohave inertia as well. Its variation is based on the tendenciesobserved in past decades and accelerated in some scenarios.
2.3. Trends of the key variables
The trends of the key variables are determined by a scenarioframework, which sets the values of the exogenous variables (orpolicies) of the model (see Fig. 2). A detailed description of thesescenarios and policies is given in Section 6.
Fig. 2. Causal loop diagram of the model with its basic elements. Scenario element
Once a scenario is set, the estimation of the energy demand iscalculated as the product of the estimated GDP (determinedexogenously) by the energy intensity. Demand is organized into 3aggregated sectors: Transportation, Electricity and IB (Industrial andBuildings) without electrical demands. Each sector's energy de-mand is generated through sectoral energy Intensities (details inSection 4). These energy demands are divided into demands ofdifferent resources following past trends: electricity from differentsources, liquid fuels, etc.
The non-renewable energy extraction (coal, oil, uranium, gas) iscompared with the demand, taking into account that it is restrictedby their maximum extraction curves (see Appendix B). The modelincludes the estimations of expansion of several technologies(renewable electricity, bioenergy, nuclear, CTL (coal-to-liquids), GTL(gas-to-liquids), etc. (as shown in Appendix C)). The policies of thedifferent scenarios determine the expansion of each technology.Finally, net CO2 emission and concentration levels to 2100 arecomputed.
s and policies are circled. IB is the acronym for Industrial and Buildings sector.
Fig. 3. Estimations of oil extraction by different authors. There is a lack of standardization in the literature. For each study, “oil” refers to only crude oil (including NGLs) [85]; crudeand unconventional [6,41,42]; crude, unconventional and refinery gains [3,123,150]; crude oil, unconventional, refinery gains and biofuels [81]; finally [16] historical data includecrude oil, shale oil and oil sands. Ref. [3] adjusts the total volume to the energy content since 1 barrel of NGL contains in reality 70% of the energy of an oil barrel.
5 <http://www.iter.org>.
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666 645
Priority is given to renewable energy (once the infrastructureis built, all the energy generated is used), and the rest of thedemand is proportionally divided between the non-renewablesources, maintaining past ratios (20-year average values fromInternational Energy Outlooks). In this way, the comparison ofdemand vs. production is done for each fuel. Since energytransitions have been shown to be slow [49], and past fuel ratiosby sectors have happened to change smoothly in the recent past[150], we consider this analysis valid in the medium term(2050).
2.4. Limitations of the model
Appendix E discusses some other omissions of the model, suchas the lack of some feedback loops, the non-inclusion of the EROEI,fuel competition or limits to minerals extraction. In this sense, wecan say that our model is conservative and its results can be seen asoptimistic. The development of a more sophisticated model withenergy-economy-climate feedbacks would be desirable, and, atpresent, the authors are oriented towards Ecological Economics inorder to apply theories that describe the real importance of naturalresources in the economic process.
Despite these simplifications, the main advantage of WoLiMis the large amount of data it integrates and its structuralsimplicity, which makes it very transparent. It is not a model thatintends to predict the future, since it only says which future isnot possible because of being not compatible with physical re-strictions. In fact, the ultimate objective of SD and scenariodevelopment is not to predict, but to understand the systemanalyzed [90,129].
The following subsections describe the energy resourcesmodeling (3.1 for non renewable and 3.2 for renewable), the energydemand estimation through sectoral intensities (Section 4) and theestimation of CO2 emission and concentration (Section 5). In themodel, we discuss the assumed potential of energy resources until2050; therefore, nuclear fusion is not considered since the ITER and
DEMO projects estimate that the first commercial fusion powerwould not be available before 2040.5
3. Energy resources modeling
3.1. Non-renewable resources
The previous model [93], extensively discusses the differentindividual fuel extraction profiles proposed in the literature. Thus,we follow their discussion and select the same profiles (updatingwhen new data is available). For some resources, we provide a “BestGuess” and “High Case” estimation based on the literature range(“Best Guess” the one considered most probable and “High Case”the one of highest resources).
3.1.1. OilThe recent estimations for conventional oil of different authors
tend to converge [29,85,128], and in order to reach stronger con-clusions, the highest estimation for conventional oil found in theliterature has been implemented (see Fig. 3).
Unconventional oil extraction is modeled following theapproach used in Ref. [33] by extrapolating the 4.5% annual growthpast trend and an optimistic 6.6% annual growth, as estimated byRefs. [53,126]. An URR of 750 Gboe is considered for non conven-tional oil after a review of other studies [6,29,54,81].
3.1.2. Natural gasIn the previous model [93], the gas profile selected was from
Ref. [81]. In this paper the recent update of the author has beenused for most scenarios [82], which assumes a combined URRfor conventional and unconventional resources of 13,000 tcf(13.6 ZJ) (Fig. 4). In some of the scenarios a higher case is taken,assuming that unconventional gas would expand significantly
Fig. 5. Estimations of coal extraction by different authors. (1 Mt ¼ 0.482 Mtoe [65]).
Fig. 4. Estimations of conventional and unconventional natural gas extraction by different authors.
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666646
more under certain favorable conditions. In these cases, the“Best Guess” of [95] that assumes almost 13,000 tcf of conven-tional gas and more than 7000 tcf (7.3 ZJ) of unconventional gasare available is used (thus 50% more than the URR estimate ofRef. [82]).
3.1.3. CoalAlthough there is a great interest in the phenomenon of peak
oil and peak gas, very few research groups work with peak coal.Even fewer consider the fact that coal is a solid mineral, which
implies different geological restrictions and different miningtechniques (Hubbert's approaches being the most used). This iswhy the chosen extraction curve is the updated to the “Highcase” estimation by Refs. [95,96] (Fig. 5), based on a miningmodel.
3.1.4. UraniumThe estimation of uranium extraction taken is the one by Ref.
[152], which includes new (more costly) ore-reserves categories, asestimated by the Nuclear Energy Association, increasing the URR
Table 1Non-renewable resources used in the model.
Resource Reference Description URR
Oil Conv. Ref. [85] high scenario. Hubbert method. 3000 Gb 16.71 ZJUnconv. Best Guess: Own projection based on Ref. [33] Extrapolation of past trends deployment (þ4.5%/yr) 750 Gb 4.2 ZJ
High case: [53] High deployment rate (þ6.6%/yr)Natural gas Best Guess: Ref. [82] Best Guess Hubbert method: “creaming curve”. 13,000 tcf 13.6 ZJ
High Case: Best Guess from Ref. [95] 12,900 tcf of conventional þ 7200 tcf of unconventional. 19,100 tcf 19.9 ZJCoal [95] High Case, static. Mining model extraction. 670 Gtoe 27.8 ZJUranium [152] Hubbert method, considering RAR (<260 $/kgU) and IR of NEA.a 19,500 KtU 8.2 ZJ
a RAR: reasonably assured resources; IR: Inferred resources; NEA: Nuclear Energy Association.
Fig. 6. Estimations of uranium extraction by different authors.
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666 647
(see Table 1) and peak production until a maximum of 100 Kt ofuranium mined is reached (Fig. 6).6
A summary of all these estimations can be found in Table 1.
3.1.5. Coal to liquids and gas to liquidsOther technologies for producing liquid fuels, such as CTL (coal-
to-liquids) and GTL (gas-to-liquids), are also considered in themodel.Different technologies exist, but all of them are characterized by lowefficiencies [63,72]. Their current production is exiguous: less than0.3 Mb/d in 2011 [150] and their growth projections from interna-tional agencies are usually relatively modest (e.g. þ11%/yr for GTL inthe New Policies Scenario of WEO [150]).
3.1.6. Integration of resources curvesTo be able to use all these data in our model we must transform
them, since it is a dynamic model that considers demand. If theworld economy goes into crisis and does not demand gas, forexample, it will not be produced. The maximum extraction curves
6 A recent paper is even more stringent, estimating that the peak will occurwithin this decade at 58 ± 4 ktons [35]. The model does not include secondaryresources of uranium (tailings, stocks and former nuclear weapons), since they willbe exhausted within a few years [35,39].
as a function of time have been transformed into maximum pro-duction curves as a function of the stock of resources (as in Fig. 7),as explained in Appendix B. Production will, therefore, be theminimum between the demand and the achievable maximumextraction.
3.2. Renewable resources
Renewable energy is usually considered as a huge abundantsource of energy; therefore, the technological limits are assumed tobe unreachable for decades, and the concern is on the economic,political or ecological constraints imposed [73]. However, severalimportant constraints limit their practical availability. In this sec-tion, we discuss the techno-ecological potential of the mainrenewable energies.7
3.2.1. BioenergyBioenergy provides approximately 10% of global primary
energy supply and is produced from a set of sources (dedicatedcrops, residues and Municipal Solid Waste (MSW), etc.) that can
7 For a detailed analysis of the modeling see Ref. [24].
Fig. 7. (a) Energy resource extraction curves as a function of time from the original references; (b) Curves of maximum extraction in function of the remaining reserves for all thenon-renewable resources (Primary Energy). The y axis represents the maximum achievable extraction rate (EJ per year) associated to the remaining reserves (EJ). For each resource,the extreme left point (that coincides with the maximum value of reserves) represents its URR. Thus, as extraction increases and the remaining reserves fall below the point wherethe maximum extraction can be achieved, the extraction is forced to decline following the estimations of the studies selected. They also show by a rhombus the 2007 level ofremaining reserves for each resource. See Appendix B for wider explanations.
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666648
serve different uses (biofuels, heat, electricity, etc.). Its techno-ecological potential estimation critically depends on the futureland availability since its land requirements are huge and theforeseeable needs of land for food and infrastructures for thegrowing population poses a limit on its expansion [30]. Thepotential of bioenergy has been established in the model basedon the land surface that could be dedicated to it. It varies be-tween occupations similar to present value (maximum100 MHa) and a maximum of 200 MHa (a complete descriptionin Appendix F).
9 We do not consider here the so called “energy trap” [99] (Zenzey, 2013). If itwere taken into account, the results would be worse (in energy terms), because theenergy needed to build the infrastructure necessary for a sustainable and renew-
3.2.2. Electrical generation from renewable resourcesThe most promising electric renewable energies are solar and
wind [124]. However, recent assessments (done by the authors ofthis paper) using a top-down methodology, which takes intoaccount real present and foreseeable efficiencies and surfaceoccupation, suggest that their potential is greatly limited bytechnical and sustainable limits [31,32].8 Thus, the evaluation ofthe global technological wind power potential, acknowledgingenergy conservation, leads to a potential of 30 EJ/yr [31]; whilethe estimation of the real and future density power of solar in-frastructures (4-10 times lower than most published studies)leads to a potential of approximately 60e120 EJ/yr [32].Following these considerations, the global techno-ecologicalpotential of electric renewable is estimated at 150 EJ (5 TWe)(see Table 3).
8 The technical potential takes into account the energy that the windmills orpanels can extract, considering current or future plausible technological effi-ciencies. Economic potential and sustainable potential are the fractions of thetechnical potential considering, respectively, the restrictions derived from the costsof the technologies and the constraints derived from sustainability and ecosystemdamage criteria (see for instance Ref. [34] for similar definitions).
In terms of investment and costs, we compute the investmentfor building new plants and to replace or re-power the alreadyexisting ones following [131], grid reinforcement costs following[94], and balancing costs as modeled by Ref. [62]9 (Table G1). For adetailed description of the modeling of electrical generation fromrenewable resources see Ref. [24].
3.2.3. Thermal renewableThe Industry and Buildings sectors are much more complex
sectors to analyze since they use all kinds of fuels and energyvectors in a great diversity of technologies. Consequently, wedecided to focus in this study on the Transport and Electricitygeneration sectors, while maintaining a high level of aggregationin IB (industrial and buildings) sectors. Thus, thermal uses ofrenewable energies (e.g. solar, geothermal) are not explicit in themodel, nor are they assigned to a concrete technology (except forthe 3rd generation biomass residues). Energy transition policiesinclude a switch to renewable, more efficient systems, as well asimprovements in construction (e.g. in order to enhance isolationand access to natural light) or even changes at a higher level (e.g.district heating), in the same way as done in WORLD3 [92]. Thesepolicies are modeled as target-policies of market penetration level
able energy system must come from current consumption of fossil fuel. Following(Zenzey, 2013, p. 80): “Unlike monetary investments, which can be made on creditand then amortized out of the income stream they produce, the energy investmentin energy infrastructure must be made upfront out of a portion of the energy usedtoday (…) The arithmetic is daunting. To avoid, for example, a 2-percent annualdecline in net energy use, replacing that loss with solar photovoltaic (with an EROEIpegged at 10:1) will require giving up 8 percent of the net energy available for theeconomy”.
Table 2Results of the sectoral energy intensity regressions for Total, Transport, and Elec-tricity generation; and of the calibration for the IB sector. All dollars in the paper arein 2011 US$. PE: Primary Energy. The World Bank database [153] is used for thehistorical series of world GDP at constant prices in US2011 T$ and Total PrimaryEnergy (PE) demand, IEA ITP [68] for Transportation PE use and US EIA [141] for theelectrical generation. IB PE intensity was calculated internally in the model for thecalibration period (1990e2010) as the subtraction of Total energy minus Transportand Electrical sector (generation and losses).
Energy sector Sectoral energyintensities
Period
Total PEdemand
Itot ¼ 0.988582Itot�1
(R2 ¼ 0.999840)EJ/US$ 1971e2010 (regression)
TransportPE demand
Itransp ¼ 0.993298Itransp�1
(R2 ¼ 0.999841)EJ/US$ 1971e2007 (regression)
Electricitygeneration
Ielec ¼ 1.00127Ielec�1
(R2 ¼ 0.999916)TWh/US$ 1980e2010 (regression)
IB PE demand IIB ¼ 0.995IIB�1 EJ/US$ 1990e2010 (calibration)
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666 649
for a given year (see Appendix D for a description of themodeling).
4. Energy demand estimation
A diversity of techniques can be used for estimating the energydemand for an economy or sector. Since the model is highlyaggregated, the Energy Intensity method, that has already beenused in similar studies [50,116] has been applied. This method issimplistic because it does not explicitly include the price and theeconomic structure; however, at medium term, energy demandand its main drivers (GDP and technological improvement) domi-nate over the variations of fuel prices [29,50,116]. In fact, prices andcosts can falsely signal decreasing scarcity. Ref. [112] demonstratesthat, when considering the size of the resource base as unknown(or ignored), it is possible to have several years of increasing
Fig. 8. Historic and estimated energy intensities by sectors. Itot refers to Total Energy Primgeneration intensity (TWh/US$). All dollars in the paper are in 2011 US$.
production simultaneous with lower prices and costs until a sud-den, intense price rise occurs with a huge cut in production, similarto the oil shock in 2007e08 [56].
Considering the sectoral Energy Intensity as energy used by asector divided by the total GDP of the economy (Eq. (1)), thismethod can be summarized as follows:
1. Estimation of the future evolution of GDP (set exogenouslydepending on the scenario),
2. Estimation of the evolution of the Energy intensity for eachsector (econometrically calculated in this study),
3. Finally, multiplying the GDP by the Energy intensity of eachsector (Ii), the Energy Demand for that sector (Ei) is obtaineddynamically, see Eq. (1):
Ei ¼ GDP$Ii (1)
Index i refers to the 3 economic sectors considered: Transport,Electric and IB (Industrial and Buildings) sectors.
The results of the sectoral historic World energy intensity re-gressions are shown in Table 2 and Fig. 8. They indicate that, in thelast 40 years, the world TP energy intensity has improved at ayearly average rate of 1.15%, but that, since 2000, its value hasremained constant at around 8 EJ/2011 US$. Transport andBuildings primary energy intensities have also improved in thelast decades, although at smaller rates (0.7% and 0.5% respec-tively). Finally, the electricity generation intensity has remainedstable at 275 TWh/2011 US$.
In order to account for the biophysical and thermo-dynamicallimits in the substitution of inputs in production in the mediumand long-term (as stated by Ecological Economics, e.g. Refs.[7,36,130]), the following expression of the energy intensity (Eq.(2)) as proposed by Ref. [118] is used:
Itot ¼ Imin þ �It¼0 � Imin
�$at (2)
ary intensity (EJ/US$), Itransp to Transportation intensity (EJ/US$), and Ielec to Electrical
Table 3CO2 emissions for non-renewable resources used in the model.
Resource Reference Value [tCO2/toe]
Coal [16] 3.96CTL [17] 6.94Natural gas Conventional [16] 2.35
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666650
AEI represents the Annual Efficiency Improvements, thus, theparameter “a”, or (1�AEI), accounts for technological change, andImin is a horizontal asymptote that represents theminimumvalue ofthe energy intensity. Both values will vary depending on the sce-nario storyline and quantification (see Section 6). The studies ofBaksi and Green (2007) and Lightfoot and Green (2002) [9,84] areused as a reference.10
5. CO2 emissions and concentrations
The model computes the CO2 emissions associated with the useof fossil fuels: coefficients from Ref. [16] for conventional and fromRefs. [17,66] for unconventional. Biofuels are far from being neutralcarbon emitters due to Indirect land use changes (ILUC); in accor-dancewith Refs. [45,55,121,154] we assign a similar emission powerto natural gas (see Table 3).
In order to assess climate change, the net11 CO2 emissions areconverted to concentration levels which assume that, in the periodstudied, the ocean and ground will continue to absorb 45% of totalemissions as in the past [23]. Due to the high inertia and long-termscope of climate change, the emission projections are extendeduntil 2100, as the IPCC usually does, with the aim of comparingconcentration levels in at least the order of magnitude.
6. Scenarios and policies of the model
As described in Section 2, the WoLiM model needs assumptionsabout the world socio-economic evolution (such as economic andpopulation growth or technological progress) as external inputs. Inorder to establish those policies in a coherent and sensible way, wehave applied the scenario methodology (e.g. IPCC's Assessmentreports [69,70,75], UNEP's Global Environmental Outlook[137,138,140] or MEA [89]).
Testing system dynamics models and obtaining results from themcanbe a cumbersome taskwhen themodels have several policies thatcan be varied at the same time. Scenario methodology offers anapproach to deal with the limited knowledge, uncertainty andcomplexity of natural and social sciences and, applied to System Dy-namics models, can be used to group the variations of policies intocoherent and meaningful scenarios. Each scenario (or storyline) rep-resents an archetypal and coherent vision of the future ewhich maybe viewed positively by somepeople and negatively by others [75,89].
By using this methodology we replicate the usual visions offuture explored by these international agencies [144], allowingthem to confront with the case of the energy development constraints.
10 A practical application for illustrating its behavior is done in Ref. [24].11 In this model version we implement the afforestation as the only CO2 seques-tration policy following [103], which analyzed the changes in the carbon cycle thatcould be achieved with a large global afforestation program covering 345 Mha.Other technologies such as CCS are not considered in this study due to their un-certain development and benefits [48,120].
In fact, to date, these international scientific bodies have largelyignored these constraints [3,64].
In this section, a summary of the most important characteristicsof the different scenario families identified in GEA studies by Ref.[144] is provided, describing first the qualitative features of eachscenario,12 and then their quantification. A Business-as-Usual sce-nario is added as reference that assumes that historical dynamicswill also guide the future).13
� Scenario 1 e Economic optimism with some market reforming:Strong focus on the mechanism of competitive, efficient market,free trade and associated rapid economic growth, but includingsome additional policy assumptions aimed at correcting somemarket failures with respect to social development, povertyalleviation or the environment. The scenario typically assumesrapid technology development and diffusion and convergenceof income levels across the world. Economic growth is assumedto coincide with low population growth (given a rapid drop infertility levels). Energy and material scarce resources areupgraded to reserves or substituted efficiently through marketsignals (price rising). Eventually, everyone will benefit fromglobalization and technological advances will remedy ecologicalproblems (e.g. ‘Environmental Kuznets Curve’).
� Scenario 2 e Global Sustainable Development: Strong orientationtowards environmental protection and reducing inequality,based on solutions found through global cooperation, lifestylechange and technology (more efficient technologies, demateri-alization of the economy, service and information economy,etc.). Central elements are a high level of environmental andsocial consciousness combined with a coherent global approachto sustainable development. Within this scenario family, it isassumed that a high level of international governmental coor-dination is necessary and possible in order to deal with inter-national problems like poverty alleviation, climate protectionand nature conservation. It entails regulation of markets but ona global scale and based on the conviction that the Earth's limitsare in sight and that therefore pro-active policies are necessary.
� Scenario 3 e Regional competition/regional markets: Scenarios inthis family assume that regions will focus more on their self-reliance, national sovereignty and regional identity, leading todiversity but also to tensions between regions and/or cultures.Countries are concerned with security and protection, empha-sizing primarily regional markets (protectionism, deglobaliza-tion) and paying little attention to common goods. Due to thesignificant reduction in technological diffusion, technologicalimprovements progress more slowly.
� Scenario 4 e Regional Sustainable Development: This scenario isthe “friendly” version of the previous one, where globalizationtends to be deconstructed and an important change in tradi-tional values and social norms happens against senselessconsumerism and disrespect for life. Citizens and countriesmust each take on the responsibilities they can bear, providingaid or setting a green example to the rest of the world, from asense of duty, out of conviction or for ethical reasons or to solveprimarily their own problems. In fact, although barriers forproducts are re-built, barriers for information tend to be elimi-nated. The focus is on finding regional solutions for current
12 We have completed the descriptions from Ref. [144] with the IPCC SRES [75]and MEA [89].13 In reality Van Vuuren et al. [144] recognize 6 scenario families. As they argue intheir paper, family scenario 1 “Economic optimism/conventional markets sce-narios” and 2 “Reformed market scenarios” are very similar. Thus, we decided tojoin them for the sake of simplicity and minimize the number of representativescenarios.
Table 4Hypothesis and policies of each scenario. Percentages refer to yearly growth rates, otherwise it is specified differently.
Scenario input 0 e BAU Scenario 1Economic optimismwith some marketreforming
Scenario 2Global sustainabledevelopment
Scenario 3Regionalcompetition
Scenario 4Regionalsustainabledevelopment
Socioeconomic(% 2010e2050)
GDPcap Hist +1.9% (1960e12) +3% +2.4% +1.1% +1.9%Population UN Medium-Variant
+0.75%+0.5% +0.65% +0.81% +0.8%
Sectoral efficiencyimprovements
aTransp Past trends (�0.67%) Rapid (�0.9%) Rapid (�0.9%) Deglobalization(�1.5%)
Deglobalization(�1.5%)
aelec Past trends (0%) Past trends (0%) Past trends (0%) Past trends (0%) Past trends (0%)aBI Past trends (�0.5%) Past trends (�0.5%) Past trends (�0.5%) Past trends (�0.5%) Past trends (�0.5%)Imin
a 25% 25% 15% 25% 15%Resource
availabilityNon-renewables Best Guess Best guess (coal, conv.
oil) Highcase (gas, unconv. oil)
Best Guess Best Guess Best Guess
CTL, GTL Crash program (+15%) Crash program (+20%) Crash program (+20%) Crash program(+15%)
Crash program(+15%)
Electricrenewables
Solar FV & CSP Medium (+15%) Past trends (+19%) Very rapid (+25%) Medium (+15%) Very rapid (+25%)Wind Medium (+20%) Past trends (+26%) Very rapid (+30%) Medium (+15%) Very rapid (+30%)Hydroelectric,Geothermal,Bioenergy & Waste
Past trends (slow) Past trends (slow) Very rapid (�3 pasttrends)
Past trends (slow) Very rapid (�3 pasttrends)
Oceanic Rapid (+20% from 2020) Rapid (+20% from 2020) Very rapid (+30% from2020)
Rapid (+20% from2020)
Very rapid (+30%from 2020)
Nuclear Constant +3% from 2015 +1.5% from 2015 Constant Progressiveshutdown
BioEnergy(100 MHa)
2nd generation Slow (+8%, 100 MHaavailable)
Rapid (+20%, 200 MHaavailable)
Rapid (+20%, 200 MHaavailable)
Slow (+8%,100 MHa available)
Medium (+15%,
3rd generation Slow (+8% from 2025) Rapid (+20% from 2025) Rapid (+20% from 2025) Slow (+8% from2035)
Medium (+15%from 2035)
Residues Slow (+8% from 2025) Rapid (+20% from 2025) Rapid (+20% from 2025) Slow (+8% from2035)
Medium (+15%from 2035)
Thermalrenewables& efficiencies
Industrial sector(market share 2050)
Low (12.1%) Medium (23.1%) Rapid (37.6%) Low (12.1%) Rapid (37.6%)
Buildings sector(market share 2050)
Low (4.7%) Medium (22.6%) Rapid (48%) Low (4.5%) Rapid (48%)
Alternativetransportb
HEV & Hybrid(market share 2050)
Medium (9%) Rapid (18%) Very rapid (36%) Medium (9%) Very rapid (50%)
NGVs Past trends (+20%annual)
Past trends (+20%annual)
Past trends (+20%annual)
Past trends (+20%annual)
Past trends (+20%annual)
Afforestation program e e 350 MHa e 350 MHa
a The minimum intensity level (Imin) is set at 25% of current intensity for scenarios BAU, 1 and 3, and at 15% for 2 and 4 following Ref. [9].b The “Alternative transport” policies are maintained while the “scarcity point” in the fuel inputs (i.e. electricity for HEV & EV and gas for NGVs is not reached).
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666 651
environmental and social problems, usually combining drasticlifestyle changes with decentralization of governance.
6.1. Quantification of the scenarios
In order to use these storylines in our model, we must set spe-cific numbers to every assumption and policy. Global14 scenarioquantification is a delicate and inherently subjective task. We havefollowed other GEA assessments as a guideline; however, di-vergences in the interpretation and hypothesis considered some-times emerge and are justified in this paragraph.
Socioeconomic inputs: GDP per capita and population evolutionestimations are taken from MEA [89]; numbers are in fact verysimilar to IPCC SRES scenarios [75].
Energy available resources: due to the Scenario 1 storyline(enhanced technological advances in extraction together with aneconomic short-term benefit priority), we consider that
14 Ethical issues regarding equitable distribution of natural capital and burdensharing rules in a resource-constrained world are beyond the scope of this paper.
unconventional oil and gas can be extracted at higher rates than forthe rest of the scenarios (“High cases” considered in Section 3). Inscenarios that prioritize the environment over the economy, weconsider that unconventional fuels are not extracted at larger levelsthan the “Best Guess” case (e.g. Refs. [106,107]).
CTL and GTL: While there is no shortage of liquids in the econ-omy, there is a growth of these technologies following recent pasttrends: strong growth for GTL and slow for CTL. When liquidshortage begins, a crash program following a logistic curve islaunched with a strong yearly growth, as indicated in Table 4. Forthe sake of simplicity, CTL deployment is not constrained, and theyearly growth is set to match current GTL growth.15 When gas and/or coal resources are not able to balance the global demand, thecrash program is stopped.
Sectoral efficiency improvements: Each sectoral efficiency(Transportation, IB and Electrical generation) is governed by itsenergy intensity (Eq. (2)) with the values in Table 4 for each
15 This can be considered as an optimistic assumption since CTL, differently toGTL, is still an immature technology (excepting South Africa) and faces significantdeployment constraints [63].
Table 5Total energy intensity evolution (average yearly reduction rates for the period2010e2050) as a result of the interpretation and quantification of scenarios and thecomparison with IPCC SRES [75] results.
Total energyintensity evolution
BAU Scenario 1 Scenario 2 Scenario 3 Scenario 4
This study �0.82% �1.04% �1.24% �0.84% �1.21%IPCC SRES [75] e �1.5% (A1) >�2% (B1) �0.65% (A2) �1% (B2)
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666652
scenario. The energy intensity of each sector (and the total energyintensity) can thus be computed when accounting for all the pol-icies (see Fig. 10 and Table 5). Electricity generation intensity ismaintained constant along the period, assuming that the currentelectrification trends will continue in all scenarios (the New Pol-icies Scenario from WEO [149] still projected 1.0 billion peoplewithout electricity by 2030). In Scenarios 3 and 4, where deglob-alization occurs, a 1.5% yearly decrease in the transport energy ef-ficiency is assumed (i.e., doubling current trends), accounting for anabsolute reduction in transport needs but also due to the promo-tion of public transport in Scenario 4.16
Notice that the evolution of all the energy intensities in ourmodel can be considered optimistic, since in recent years, thesectoral energy intensities yearly averages have diminished atsmaller rates than historical values, the total energy intensity hasbeen constant since 2000, and electricity generation intensity haseven increased (Fig. 8).
Nuclear: Considering the study by Schneider et al. [119] of thenuclear status in the world, the conservation of the already existingpower in the coming years would already be an optimisticassumption; we assign this case to BAU and Scenario 3. For Sce-narios 1 and 2, where nuclear may be promoted, we take asreference the World Nuclear Association (WNA) forecast of 1e2%/year growth for the coming decades [35]. Finally, for scenario 4,where the environment is actively protected, we program a pro-gressive phase-out as nuclear power stations reach the end of theirlifetime.
Electric/Hybrid (HEV&hybrid) car: the evolution of hybrid andelectric vehicles in our model follows the estimations of EVI, whichis “a multi-government policy forum dedicated to accelerating theintroduction and adoption of electric vehicles worldwide” thatseeks to “facilitate the global deployment of at least 20 millionpassenger car EVs by 2020” [38]. We will consider this forecast asan optimistic development and we assign it to scenarios 2 and 3,while for BAU and Scenario 3 wewill keep half of the projection. ForScenario 4, we will interpret the “lifestyle change” as a higher shift.After 2020, the growth rate is assumed to increase two-fold for allscenarios, assuming that a shift to alternative mobility systems will,in any case, be promoted, due to the scarcity of conventional liquidfuels in all scenarios.
Natural Gas Vehicles: Despite the strong growth in the pastdecade (þ20% per year), the total number of 16.7 million NGVs(http://www.iangv.org/current-ngv-stats/) still pales in compari-son to a total worldwide number of around 1150 million motorvehicles in 2009 [153] e i.e.1.45% of the total. Thus, due to theinsensitivity of the model to different NGV growth rates (because of
16 Potential reductions of energy consumption in the Transportation sector in thedeglobalization scenarios are in fact very limited due to the small contribution ofworld aviation and marine bunkers in relation to the primary energy used by thewhole sector (below 13%, while road transport accounts for more than 65% [68]). Infact, a wide range of deglobalization scenarios can be conceived, from (world)regionalization of the exchanges to more local reconfigurations that would unfoldin very different energy use patterns (e.g. Ref. [21]).
reaching the gas peak), for the sake of simplicity, an annual growthrate is assumed for all scenarios that follows the past trends.
Bioenergy: As stated in Section 3.2, a very large surface dedicatedto bioenergy at a global level is not compatible with future sce-narios such as the ones explored in this paper. As a reference, sincethe year 2000, the area from southern countries that has beenbought or long-term rented by transnationals and investmentfunds has been estimated at more than 80 MHa [4]. Two possibil-ities of bioenergy expansion are considered. For scenarios 3 and 4(regionalization), land grabbing is not going to increase signifi-cantly from present levels in the future (maximum 100 MHa), sincecurrently, there is a worldwide rush for land (around 1.7% of agri-cultural area has been reported to have been bought or rented forlong periods of time since the year 2000, [4]). For scenarios 1 and 2,a “south occupation” that would be deployed in a maximum of200 MHa is assumed (more details in Appendix F and Ref. [24]).
Renewable and efficiency improvements in the Industrial andBuilding sectors: different levels of penetration by renewable tech-nologies to 2050 are considered. More potential is assigned toBuildings than to Industry (e.g. Ref. [37]).
Carbon-climate policies: Storylines from scenarios BAU and 3exclude the adoption of carbon valuation. The scenario 1 storylinesuggests that measures could be taken, but probably too late. Thus,effective carbon valuation only seems probable in scenarios 2 and 4with proactive environmental protection. Although carbon valua-tion would intensify the transition to a low carbon economy, weconsider that many of the changes it would induce are in factalready implicit in the interpretation of both storylines. Addition-ally, we consider that a world afforestation program is set from2020 as proposed by Ref. [103]. They analyzed the changes in thecarbon cycle that could be achieved with a large-scale global pro-gram covering 350 MHa. Thus, a maximum carbon capture of1.5 GtC/year, 50 years after the start of the program, would beattained. Other technologies such as CCS are not considered in thisstudy due to their uncertain development and benefits [48,120].
Table 4 provides a detailed summary of the policies imple-mented for each scenario.17
7. Results and discussion
In this section the results of our model to 2050 under the sce-narios described in Section 6 are presented. It should be recall thatsome important issues have not been integrated in the modeling(see Appendix E), most of these issues have the potential to worsenthe results obtained.
A fundamental consideration must be made: as the model doesnot integrate a feedback between energy scarcity and GDP, if thedemand cannot be fulfilled, a divergence appears between demandand energy supply (though in the real world there would be aninteraction that would reduce this gap). We will qualitativelyinterpret that small divergences are compatible with the storyline;however, large divergences will be interpreted as potential energy-scarcity challenges that would make the scenario unfeasible.
7.1. Socioeconomic inputs and sectoral efficiencies
Socioeconomic inputs considered for each scenario are pre-sented in Fig. 9. Population increase is similar in all scenarios due toits high inertia and estimates vary between 8.3 (Scenario 1) and 9.5billion people in 2050 (Scenarios 3 and 4). Scenarios are morediverse in terms of GDP: in 2050, Scenario 1 almost doubles the
17 In fact, although IPCC SRES scenarios [75] do not consider explicitly the use ofpolicies, it has been argued that they are implicitly assumed [5,52,108].
Fig. 10. Total and by sector Primary Energy (PE) intensity evolution for each scenario in EJ/2011 US$. TPE: Total Primary Energy, IB: Industrial and Buildings sectors. Historic datafrom Refs. [68,153] and our own adjustments for the IB and Electricity PE intensity past evolution as explained in Section 4.
Fig. 9. Socioeconomic exogenous inputs for each scenario: (a) GDP (2011 US$) and (b) Population. Historic data from World Bank database [153]. Since in this model version nofeedbacks are applied, impacts of energy supply scarcity and climate change are not computed in the projections and these variables are thus not modified.
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666 653
GDP estimated in Scenario 3 (which doubles the 2010 value). Also,the variety of policies and different sectoral efficiency consideredunfold in very different (always decreasing) sectoral energy in-tensity paths (Fig. 10).
In Table 5, the total energy intensity yearly average decreaseobtained for each scenario is represented and compared with thedeclines assumed by IPCC SRES [75]. Our results are in the range0.8%e1.25% and coincide with Refs. [9,84,108] that find greater ef-ficiency improvements implausible due to the biophysical limits inprocess substitution, as proposed by A1 and B1 SRES scenarios [75].
7.2. Results by sector and fuel
7.2.1. Electricity generationThe comparison of the electricity generation and demand from
different sources can be seen in Fig. 11. In the scenarios where therenewable technologies are promoted at a very rapid pace (2 and 4),
electricity supply is roughly able to fulfill the demand, but in sce-narios BAU and 1, renewable electricity cannot sustain theincreasing demand by 2030. In Scenario 3, even the smallestgrowth of the demand cannot be compensated because of themodest growth of renewable technologies.
Wind maximum potential is reached in the 2030s, and solargrowth slows down significantly by 2050 due to the proximity of itsmaximumpotential. Uranium restrictionsmake nuclear technologylargely irrelevant.
However, the massive expansion of renewable technologies hasrepercussions. Fig. 12a shows the proportion of variable electricgeneration technologies (wind and solar) in function of the totalproduction. In Scenarios 2 and 4, this proportion exceeds 50% of thetotal generation by 2050, which would imply an important chal-lenge for the integration of intermittent production [133]. In termsof investment (Fig.12b), electric renewable deployment investmentwould remain below 1.5% of the total world GDP of all scenarios and
Fig. 11. Electricity generation and demand (TWh/yr) by fuel source for each scenario. Historic consumption data increased by transportation losses (9% average) are taken from USEIA db [141].
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666654
is in the samemagnitude order as other studies (e.g. Bloomberg NewEnergy Finance18 or the Energy [R]evolution [131]).
7.2.2. TransportationThe comparison of the demand and the supply of energy for
transportation can be seen in Fig. 13. In spite of the diversity ofpolicies applied in the different scenarios, the peak of conven-tional oil in the early 2010s determines a decline or stabilization ofenergy available for transportation. Biofuels, alternative electric,
hybrid and gas transport, CTL>L (that does not developsignificantly in any scenario due to the ending of the crash pro-grams when gas and coal reach their peaks), efficiency improve-ments, and even the higher development of unconventional oil inScenario 1, cannot reach a substitution rate able to compensatethe conventional oil decline. Thus, as also found in Ref. [93], en-ergy shortages appear in the Transportation sector for all scenariosbefore 2020 (Fig. 15).
7.2.3. Total Primary Energy extractionThe comparison of the Total Primary Energy (TPE) demand and
extraction can be seen in Fig. 14. Broadly speaking, TPE extractionremains below 800 EJ/yr in 2050 (around þ50% in relation to the
Fig. 12. For each scenario, (a) Proportion of variable electric generation (wind and solar) vs. total and (b) Proportion of the investment in electric renewable related to total GDP.
19 Institutional talks and agencies consider that the critical threshold for stabi-lizing climate change “at a level that would prevent dangerous anthropogenic inter-ference with the climate system” is 450 ppm: e.g.: UNFCCC (Cancun Agreements), UE,IEA (450 ppm Scenario). However, even this higher value would be exceeded by2050 in all scenarios.
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666 655
2010 level). Moreover, the past growth trend (þ2.6%/yr 1965e2012[16]) cannot be maintained and the yearly energy available by 2050is either decreasing (�0.7%/yr in scenarios BAU and 3), roughlystabilized (slower growth than 0.5%/yr in Scenarios 1 and 4) orgrowing slightly at around 1%/yr in Scenario 2). This occurs becausethe decrease in fossil fuel extraction can only be partiallycompensated by renewable energies, alternative policies and effi-ciency improvements. In fact, between 2020 and 2030, differencesbetween supply and demand appear to be significant in all sce-narios (Fig. 15).
7.2.4. Comprehensive analysis of scenarios: energy scarcity matrixIn order to analyze the supply constraints on the demand of
each sector and energy resources for all scenarios, we represent the“Energy Scarcity Matrix” in Fig. 15 and Table 6. For each economicsector and non-renewable resource, each point represents the datewhen the relative difference between the demand and supply isgreater than 5%.We select 5% as a qualitative threshold as it is whenthe price-mechanism adaptation could force important socio-economic structural changes that would modify the underlyinghypothesis of the scenarios and the model. For renewable re-sources, each point represents the datewhen 95% of the potential isreached. A similar sequence of facts appears for all scenarios:
1. Liquid scarcity in 2015e20 precipitates energy scarcity in theTransportation sector immediately afterwards, and in the IBsector a few years later.
2. Total Primary Energy and Gas scarcity roughly coincide in2020e25.
3. By 2035, Coal supply is not able to cover its demand in anyscenario. Restrictions in the coal supply could appear soonerthan usually expected.
4. Electricity generation for all scenarios is not able to fulfill thedemand in 2025e2035 (in spite of the strong promotion ofrenewable energies in some of them).
5. Uranium resources are able to provide the mineral needed tomaintain the current production to 2050; however, when even amodest increase in capacity is considered, uranium extractionlimits appear.
6. A large expansion of electric renewable energies move us closeto their potential limit (e.g. solar), which may even be reachedbefore 2050 (wind and hydroelectric).
As revealed by the scenario approach, these dynamics are notindependent: when increasing the number and intensity of linksbetween the non-renewable energies (transition policies), thedifferent peaks tend to converge in time.
Finally, in Fig. 16a, the evolution of the extraction of fossil fuelsfor all scenarios is represented. In spite of the diversity of policiesand assumptions applied (notably renewable development, seeFig. 16b), a “decline path” for the future extraction of fossil re-sources emerges, reaching a “plateau” at around 500e525 EJ ofmaximum extraction between 2020 and 2035, depending on thescenario. This plateau is followed by a sharp reduction between 1and 1.5% per year.
7.3. Global warming
The results of CO2 for all scenarios show a peak of emissions in2020e2030 at 40e45 GtCO2, which is a value 35e50% higher than2005 emissions (Fig. 17a). Scenarios BAU, 2, 3 and 4 project adecline along the century in a path very similar to B1 from IPCCSRES [75]. Scenario 1, however, maintains higher emission levels,similar to the A2 SRES scenario due to the higher extraction ofunconventional fuels.
Likewise, all scenarios project similar concentration values inthe first half of the century (Fig. 17b), reaching around 475 ppm in2050. In 2100, all scenarios reach the 550 ppm level, and Scenario 1almost reaches 600 ppm. Paleoclimate evidence and ongoingclimate change suggest that CO2 would need to be reduced to atmost 350 ppm if humanity wishes to preserve a planet similar tothat on which civilization developed and to which life on Earth isadapted [57,58,114].19 In fact, if high concentration levels aremaintained during a certain time, anthropogenic climate changecould be boosted by (irreversible) slow feedback dynamics (e.g. ice-free poles). Thus, in all of our scenarios, despite the fact that CO2emissions fall because of the peak of fossil fuels, these concentra-tions during the 21st Century are highly alarming and dangerous.Moreover, we consider our results to be optimistic, since the
Fig. 13. Transportation Primary Energy demand and supply by source fuel (EJ/yr) for each scenario. Historic consumption data is taken from Ref. [68]. Other liquids include un-conventional oil, CTL, GTL and refinery gains. Note: Primary Energy demand is dynamically corrected to take into account the fact that renewable technologies are more efficientthan fossil-fuel based ones.
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666656
absorption capacity of natural sinks is likely to decrease as theplanet warms [23,83].
Some other research teams that have studied the links be-tween the structural limits to fossil fuel supply and ClimateChange [18,102,145] and have found that the emissions levels tendto align with low and medium scenarios from IPCC [64]. Our re-sults coincide with them; however, even these relatively “low”
emission profiles imply that climate change could reachdangerous dimensions.
7.4. Summary and discussion on the results
Our results confirm the short-term lock-in of energy de-velopments and suggest that, the world socioeconomic system will
not be able to follow any of the scenarios proposed to 2050.Specifically:
- Electricity generation seems to be the least worrisome sector,especially in scenarios where electrical renewable generation isstrongly promoted. In such cases, saturation in their expansionpotential by 2050 is appreciable for some technologies (e.g.hydro and wind).
- Transportation: all of the scenarios presented are unfeasiblebefore 2020. Biofuels, alternative electric, hybrid and gastransport, CTL>L, efficiency improvements and even thehigher development of unconventional oil cannot reach a sub-stitution rate able to compensate the conventional oil decline(even taking the highest estimations for oil resources). In
Fig. 14. Total Primary Energy extraction and demand by source fuel (EJ/yr) for each scenario. Note: Primary Energy demand is dynamically corrected to take into account the factthat renewable technologies are more efficient than fossil-fuel based ones.
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666 657
scenarios where deglobalization is simulated, the supply-demand gap is strongly reduced due to a decrease in energyrequirements. We identify this feature as a “clue” to developingsustainable scenarios.
- Total Primary Energy (TPE) extraction: the supply of TPE can bestabilized or even grow in some scenarios to 2050, but the pastgrowth trend (þ2.6%/yr 1965e2012) cannot be maintained,since the decrease in fossil fuels extraction can only be partiallycompensated by alternative technologies and efficiencyimprovements.
- Emissions: All scenarios show a peak in CO2 emissions in2020e2030 at 40e45 GtCO2. These emissions profiles are lowerthan high-medium emissions scenarios from the IPCC; however,they already have the potential to lead to a climate change ofdangerous dimensions.
Our results do not intend to describe a plan concerning how theglobal system will evolve, since, once the first disequilibrium isreached in a sector, the systemwould evolve in a different way fromthe scenario proposal. However, compelling conclusions can be
Fig. 15. Energy Scarcity Matrix of the scenarios. For each non-renewable resource/sector, we mark the “scarcity point” when the relative difference between supply and demand isgreater than 5%. For renewable energies, each point represents the date when 95% of the potential is reached. Strong similarities in the relative scarcity outcomes between scenario1 and BAU are evident. Note: It may happen that a given resource does not reach its “scarcity point” for a specific scenario (e.g. solar in scenarios BAU, 1 and 3).
Table 6Supply-demand divergence (5%) and potential reached (95%) range in the 5modeled scenarios for all fuels and sectors. Data from the Energy Scarcity Matrixfrom Fig. 15.
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666658
extracted: the results indicate that from the current decade, theworld socioeconomic system will progressively be forced to becomeindependent of cheap and abundant energy resources that wereavailable in the past. In fact, the data of the last few years suggeststhat past trends are changing (the energy consumption of OECD
Fig. 16. Primary energy resources extracted by scen
countries is falling, the oil consumption of Southern Europe isdropping while suffering severe economic crisis, etc.).
We recall again that the exclusion of important issues in the model(see Appendix E) could only exacerbate these trends. Thus, the ob-tained results suggest that the current socioeconomic paradigmmay not be sustainable and continuous economic growth may berather more the problem than the solution. However, GDP was notdesigned to measure social welfare (e.g. Ref. [143]); and researchwith welfare indicators show not only that, above a certain level,there is no link between higher GDP per capita and subjectivewellbeing, but reductions in GDP per capita may be welfareenhancing [80]. Thus, different socio-economic paradigm scenarioscan be modeled in order to propose real sustainable future pathssuch as those proposed by Degrowth [78], Steady-State [27,79] orthe New Economics of Prosperity [76] and are the subject of currentresearch.
We add two considerations in relation to the modelingassumptions:
- Of the fossil fuels, coal is the most abundant. However, it is alsothe least studied from a depletion point of view. Thus, in order toreduce the uncertainty in future global studies, wemake a call tomotivate further research on this topic.
ario: (a) Fossil fuel and (b) Renewable energies.
Fig. 17. a) Evolution of net CO2 emissions (Gt CO2) for each scenario and comparison with the SRES scenarios A1, A2, B1 and B2; b) Evolution of net CO2 concentration (ppm) for eachscenario during 21st Century compared to past historical observations at Mauna Loa and pre-industrial value. Horizontal lines represent the pre-industrial CO2 concentration value(solid) and two different representative thresholds in the literature: 350 ppm and 450 ppm (dashed).
20 For a complete description of the model, please see Ref. [24]. <http://www.eis.uva.es/energiasostenible/?page_id¼2056>.
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666 659
- All GEA scenarios consider GDP growth. However, our resultssuggest that the objective of continued exponential GDP growthin a system incapable of effective energy-material decouplingshould be promptly replaced, instead, by the objective ofimproving economic welfare.
8. Conclusions
In this paper we introduce and apply a System Dynamics model,WoLiM, which aims to fill a gap found among Energy-Economy-Environment models, since few of them integrate the estimationsof fossil fuel depletion and alternative energy expectations with theenergy demand generated by the socio-economic system. Themodel is applied to a set of scenarios that replicate the habitualscenarios in Global Environmental Assessment studies [144], IPCC'sAssessment reports [69,70,75], UNEP's Global EnvironmentalOutlook [137,138,140] or MEA [89].
The results show that a significant systemic-energy scarcity riskexists: future global energy demand-driven transitions as performed inthe past might be unfeasible. These critical energy constraints havethe potential to provoke unexpected abrupt changes in societiesand the world configuration, making all 5 implemented families ofscenarios from the GEA studies impossible to achieve by 2050.Transportation is the most critical sector due to the stagnation ofliquids fuel production and the inefficacy of all compensation pol-icies before 2020. The Electricity sector seems the least worrisome,especially in scenarios where electrical renewable generation isstrongly promoted. However, CO2 levels still have the potential tolead to a climate change of dangerous dimensions by the mid-century. Moreover, the use of unconventional fuels in a context ofrising demand-supply divergence will tend to induce energy pricesto grow in the future with very likely adverse economic impacts[99,136].
In order to find global scenarios compatible with fossil fuel re-strictions and sensible limits to technological development, we areobliged to set hypotheses which are hardly used in Global Assess-ment scenarios, such as zero or negative economic growth.Therefore, an authentic economic paradigm shiftmight be needed inorder to avoid dangerous energy lock-in pathways in a context ofclimate deterioration in the coming decades.
Most of the current Economy-Energy-Environment models tendto use (very) large resource estimates that are subject to high un-certainties and are strongly biased towards overestimation. The
analysis performed here shows that depletion should be incorpo-rated into such policy-influential analyses as the IEA and IPCCreports.
In spite of our narrower scope, our conclusions meet with theLimits to Growth reports (reinforced after 40 years of favorablecomparison [135,134]), which stated that “current policies willproduce global overshoot and collapse through ineffective efforts toanticipate and cope with ecological limits” in the first half ofthe 21st Century, pledging “profound, proactive, societal innova-tion through technological, cultural and institutional change”[90e92].
Disclaimer
The opinions expressed in this paper are the authors' ownopinions and do not necessarily correspond with those of the LowCarbon Programme.
Acknowledgments
The authors gratefully acknowledge Steve Mohr and GaetanoMaggio for their valuable comments and share of data. This workhas been developed within the project CGL2009-14268 funded bythe Spanish Ministry of Science and Innovation (MICINN). Addi-tionally, I~nigo Capell�an-P�erez wishes to thank the Consejería deEconomía y Empleo of la Junta de Castilla y Le�on (Programa deFormaci�on mediante pr�acticas en materia de investigaci�on einnovaci�on tecnol�ogica) and the REPSOL Foundation for the supportthrough the Low Carbon Programme (www.lowcarbonprogramme.org).
Appendix A. Basic structure of the model
Figs. A1 and A2 show an overview of the Forrester diagram ofWoLiM, where the main relationships and subsystems can be seen.Demands are shown in green, non renewable resources in lightblue, renewable electricity in dark blue, policies in red and emis-sions in orange.20
Fig. A1. Forrester diagram of WoLiM model (left side). Stocks are represented as squares, flows by the arrows related to stocks, variables are represented by circles and constants byrhombus. Most of the relationships between variables are represented by lines but some are hidden for simplicity.
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666660
Fig. A2. Forrester diagram of WoLiM model (right side). Stocks are represented as squares, flows by the arrows related to stocks, variables are represented by circles and constant byrhombus. Most of the relationships between variables are represented by lines but some are hidden in order to simplify the graph.
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666 661
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666662
Appendix B. Integration of resource curves
The maximum energy resource extraction curves, described inSection 3, are curves of maximum energy extraction as a function oftime. In order to use them in themodel they have been transformedinto maximum production curves as a function of resources.
In these curves, as long as the resources are large, extractionwillnot be limited physically and we make it equal to the totalmaximum production. When the resources diminish, physicallimits start to appear and production is reduced. In this way, themodel uses a stock of resources and studies how this stock isemptied depending on production, which is in turn determined bydemand and maximum extraction.
Fig. B1 gives a hypothetical example of the dynamic model used(left) and an example of a maximum production curve (right). Thex-axis of Fig. B1 (right) represents the stock of non-renewable en-ergy available. The y-axis represents the maximum production ofthis energy that could be obtained depending on the stock of theresource still unexploited. As can be seen, when the resourcesdiminish, the maximum production decreases until it reaches zero(when the resource is exhausted). The Forrester diagram of Fig. B1(left) shows the stock of resources. A variable called maximum pro-duction is calculated as a function of the stock of resources and acurve similar to the one of Fig. B1 (right). Stock of resources isemptied by the flow called Extraction, whose value is the minimumbetween demand and maximum production.
Fig. B1. Maximum production curves as a function of resources. Left: the Forrester diagram used to model extraction. Right: a curve of maximum production (solid) compared withthe demand (dashed). Both curves meet when the peak of the resource is reached.
Appendix C. Renewable energies modeling
The growth of the renewable electricity production from allsources is modeled by a similar structure to the one presented inFig. C1 for solar. The Forrester diagram shows the stock of renew-able electricity infrastructure (solar_TWe) with its two flows: theinflow of new infrastructure determined by investments (new_s-olar_TWe), and the outflow determined by the depreciation(depreciation_solar) driven by the lifetime (life_time_solar).
Fig. C1. Structure of the renewable electric technologies (here, solar).
Therefore, the equation that determines Solar_TWe is:
dðsolarTWeÞdt
¼ new_solar_TWe� depreciation_solar (D2)
Replacement_solar just compensates for the depreciation rate,and P1_solar represents the annual growth considered in eachscenario. However, this growth is adjusted to a function that in-troduces diminishing returns on the new solar power (new_s-olar_TWe) depending on the proximity to the potential(max_solar_TWe); creating a feedback loop that reduces theexogenous growth initially set (logistic growth):
The model also accounts for the electrical production (solar_-production_TWh), the land occupied (surface_MHa_solar) and theinvestment required (invest_solar_Tdolar).
Appendix D. Modeling of alternative technologies and savingpolicies
The policies that represent alternatives to oil, non-electricalrenewable energies and saving (biofuels, electric and hybridvehicle, train, savings and renewable thermal energy for buildingsand industry) are described in the model with a similar structure tothe one represented in Fig. D1 (savings in the industry sector in thiscase). The thermal uses of renewable energies are not explicit orassigned to a concrete technology (except for the 3rd generationbiomass residues), but modeled as a general policy, in the samewayas done in WORLD3 [92].
In the example of Fig. D1, the total Industrial energy demand(Industry_EJ) is calculated in a different part of the model (as afunction of GDP and sectoral intensities). The stock variable per-cent_saving_I represents the share to the total Industrial energy de-mand that is concernedwith the transition policies. This variable is astock because it is assumed that these savings accumulate as thechange to better equipment is done. The variable percent_saving_Icauses a drop in energy demand, and the variables dem_indus-trial_after_savings and I_after_renew account for the new demand,which is divided into the demands of individual fuels (I_gas, I_coal,I_oil, I_renew)according toa share consistentwith thepastevolution.
Fig. D1. Forrester diagram of the representation of the Industrial sector and the pol-icies applied.
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666 663
Appendix E. Limitations of the modeling
As mentioned in the paper, the modeling of complex systemsalways implies a trade-off between simplicity and the loss of detail.Thus, uncertainties and limitations arise: some are solvable (andare targeted as “future research directions”), while others arerelated to unavoidable judgment calls in the extrapolation of thefuture. Among the first are:
- Non-inclusion of Energy Return of Energy Investment (EROEI): Themodel operates in terms of primary energy, but in reality theuseful energy used by society (Net Energy) in the future maydecrease at the same time as the EROEI of the non-renewableresources diminishes, due to the smaller EROEI of unconven-tional resources [98]. Some modern renewable energies alsoperform low EROEI ratios (e.g. solar [110]).
- Non-inclusion of material limits and other non energetic renewablesources (e.g. water availability [109], minerals (e.g. phosphorus[25], copper [59]).
- Absence of dynamic feedback between the main subsystems. In thismodel version, climate impacts and energy scarcity are not fedback to the economic system. Similar studies have shown thatmodels are biased optimistically when feedback is omitted (e.g.Refs. [11,111]). The MEA [89] report concluded that approxi-mately 60% (15 out of 24) of the ecosystem services examinedare currently being degraded or used unsustainably. Also,
Table F1Bioenergy power density and potentials assumed for each resource. Other potential resuncertainties of the technology and the long-term nature of its eventual commercial app
Reference
2nd generation Marginal lands [47]World average [30]
3rd generation(from 2025)
Dedicated crops [146]Agriculture & Forestry residues Own estimation
NPP: Net Primary Production.a Field et al. [47] find that 27 EJ of NPP can be extracted from 386 Mha of marginal lab The gross power density for the best quality lands was estimated at 0.3e0.36 W/m2
Ref. [114] identifies 3 out of 10 planetary boundaries that havealready been overstepped. However, high uncertainties areinvolved in the feedback quantification and remain beyond thescope of this paper.
- Others: intermittency of renewable energies, non-considerationof phenomenons such as the “energy trap”, the rebound effect,conflicts (within and between countries, e.g. corruption, wars),unexpected events (e.g. natural disasters), etc.
The omission of restrictionswhen solving a system can only leadto optimistic results. However, interesting conclusions have alreadybeen extracted and ongoing research on these issues will explorethe influence of these constraints.
On the other hand, other assumptions such as the non-modelingof technology-fuel competition (through cost and efficiency astypically done in demand-driven models), might seem as in sig-nificant weakness of the model. However, since in all scenarios thepeak of all fossil fuels occurs in the range of 15e20 years, theintroduction of the competition would only tend to slightly delaythe first “scarcity points” while hastening the last ones. In brief, foreach scenario, the points in Fig. 15 would tend to converge in time,thus, not affecting themain conclusions of thework. However, froma societal point of view, the transition might be less challenging ifthe “scarcity points” are more spread in time.
Appendix F. Potential of bioenergy
The techno-ecological potential estimation of bioenergy de-pends critically on the future land availability. The foreseeablegrowth of land for food over the next few decades (due to popu-lation and affluence growth) is projected to be 200e750 MHa[10,20,113,117], while the projected growth of new infrastructuresbecause of population and affluence growth is more than 100 MHa.Moreover, it is estimated that current and future crop yields will beaffected negatively by climate change [74], offsetting potentialproductivity gains from technological innovation. According toRef. [44], there were 1526 MHa of arable land and permanent cropsin 2011. In view of the current situation, in which almost 15% of theworld population is undernourished [43], a very large surface forbioenergy at global level is not compatible with future scenarios,such as the ones explored in this paper.
For the sake of simplicity, we decided to divide it into 3 cate-gories for differentiated uses: traditional biomass, dedicated cropsfor biofuels and residues for thermal uses (Municipal Solid Wasteand 3rd generation). The techno-ecological potential estimation ofthese categories is a sensitive and complex task: different lands(e.g. current arable vs. marginal) have different productivities, landcompetition issues, etc. The energy density and potentials assumedfor each resource are presented in Table F1. These values are basedon estimations from Refs. [30,47,139,146] and our assumptions aredetailed in Ref. [24].
ources, such as 4th generation biomass (algae), are not considered due to the highearance [77].
nds. A transformation efficiency to biofuels of 15% is assumed.in Brazil [30].
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666664
Appendix G. Potential of renewable electricity
Techno-ecological potential of renewable energies as estimatedby Refs. [24,31,32] are shown in Table G1. These limits are lowerthan some other estimations found in the literature mainly becausethey consider aspects frequently ignored such as the prorateddegradation of the cells over the entire life cycle, maintenance, self-consumption or the real land occupation of the solar parks (notonly he panels).
Table G1Data of electric renewable in the model. “TWe” represents power electric produc-tion: TWh/8760.
a The learning curve for wind is adapted from Ref. [131] in order to aggregate bothonshore and offshore wind.
b The solar investment cost is maintained constant after 2030 since we judge it tobe too optimistic that the solar technologies will manage to be less expensive thanwind. In fact, in recent years, the price of solarmodules has fallen significantly due toefficiency improvements but also to dumping and excess capacity effects in thecrisis.
References
[1] Adelman MA. Mineral depletion, with special reference to petroleum. RevEcon Stat 1990;72:1. http://dx.doi.org/10.2307/2109733.
[2] Aleklett K. Reserve driven forecasts for oil, gas and coal and limits in carbondioxide emissions (Discussion Paper No. 2007-18). OECD InternationalTransport Forum; 2007.
[3] Aleklett K, H€o€ok M, Jakobsson K, Lardelli M, Snowden S, S€oderbergh B. Thepeak of the oil age e analyzing the world oil production reference scenario inWorld Energy Outlook 2008. Energy Policy 2010;38:1398e414. http://dx.doi.org/10.1016/j.enpol.2009.11.021.
[4] Anseeuw W, Boche M, Breu T, Giger M, Lay J, Messerli P, et al. Transnationalland deals for agriculture in the global south. 2012.
[5] Arvesen A, Bright RM, Hertwich EG. Considering only first-order effects?How simplifications lead to unrealistic technology optimism in climatechange mitigation. Energy Policy 2011;39:7448e54. http://dx.doi.org/10.1016/j.enpol.2011.09.013.
[6] ASPO. ASPO Newsletter n. 100; 2009.[7] Ayres RU. On the practical limits to substitution. Ecol Econ 2007;61:115e28.
http://dx.doi.org/10.1016/j.ecolecon.2006.02.011.[8] Ayres RU, van den Bergh JCJM, Lindenberger D, Warr B. The underestimated
contribution of energy to economic growth. Struct Change Econ Dyn2013;27:79e88. http://dx.doi.org/10.1016/j.strueco.2013.07.004.
[9] Baksi S, Green C. Calculating economy-wide energy intensity decline rate:the role of sectoral output and energy shares. Energy Policy 2007;35:6457e66. http://dx.doi.org/10.1016/j.enpol.2007.08.018.
[10] Balmford A, Green RE, Scharlemann JPW. Sparing land for nature: exploringthe potential impact of changes in agricultural yield on the area needed forcrop production. Glob Change Biol 2005;11:1594e605. http://dx.doi.org/10.1111/j.1365-2486.2005.001035.x.
[11] Barney GO. Global 2000 report to the President of the United States: thesummary report e with environment projections and the Government'sGlobal Model. Entering the 21st Century, vol. 1. Pergamon Press; 1980.
[12] Bassi AM, Shilling JD. Informing the US Energy Policy debate with threshold21. Technol Forecast Soc Change 2010;77:396e410. http://dx.doi.org/10.1016/j.techfore.2009.10.007.
[13] Benes J, ChauvetM, Kamenik O, KumhofM, LaxtonD,Mursula S, et al. The futureof oil: geology versus technology (No. WP/12/109). IMF Working Paper. 2012.
[14] Bithas K, Kalimeris P. Re-estimating the decoupling effect: Is there an actualtransition towards a less energy-intensive economy? Energy 2013;51:78e84. http://dx.doi.org/10.1016/j.energy.2012.11.033.
[15] Bouwman A, Kram T, Goldewijk KK. Integrated modelling of global envi-ronmental change: an overview of IMAGE 2.4. Bilthoven: NetherlandsEnvironmental Assessment Agency; 2006.
[16] BP. BP statistical review of World Energy June 2013. Statistical Review ofWorld Energy. British Petroleum; 2013.
[17] Brandt AR, Farrell AE. Scraping the bottom of the barrel: greenhouse gasemission consequences of a transition to low-quality and synthetic petro-leum resources. Clim Change 2007;84:241e63. http://dx.doi.org/10.1007/s10584-007-9275-y.
[18] Brecha RJ. Emission scenarios in the face of fossil-fuel peaking. Energy Policy2008;36:3492e504. http://dx.doi.org/10.1016/j.enpol.2008.05.023.
[19] Brown JH, Burnside WR, Davidson AD, Delong JR, Dunn WC, Hamilton MJ,et al. Energetic limits to economic growth. BioScience 2011;61:19e26.http://dx.doi.org/10.1525/bio.2011.61.1.7.
[20] Bruinsma J. World agriculture: towards 2015/2030: an FAO perspective.Earthscan/James & James; 2003.
[21] Bueno G. Analysis of scenarios for the reduction of energy consumption andGHG emissions in transport in the Basque Country. Renew Sustain EnergyRev 2012;16:1988e98. http://dx.doi.org/10.1016/j.rser.2012.01.004.
[22] Campbell CJ, Laherr�ere J. The end of cheap oil. Scientific American; 1998.[23] Canadell JG, Qu�er�e CL, Raupach MR, Field CB, Buitenhuis ET, Ciais P, et al.
Contributions to accelerating atmospheric CO2 growth from economic ac-tivity, carbon intensity, and efficiency of natural sinks. Proc Natl Acad Sci USA2007;104:18866e70. http://dx.doi.org/10.1073/pnas.0702737104.
[24] Capell�an-P�erez I, Mediavilla M, de Castro C, Miguel LJ. World limits model(WoLiM) 1.0-model documentation. Technical Report. Spain: Energy andSystem Dynamics Group of the University of Valladolid; 2014., http://www.eis.uva.es/energiasostenible/?page_id¼2056&lang¼en.
[25] Cordell D, Drangert J-O, White S. The story of phosphorus: global food se-curity and food for thought. Glob Environ Change 2009;19:292e305. http://dx.doi.org/10.1016/j.gloenvcha.2008.10.009.
[26] Dale M. Meta-analysis of non-renewable energy resource estimates. EnergyPolicy 2012;43:102e22. http://dx.doi.org/10.1016/j.enpol.2011.12.039.
[27] Daly HE. Beyond growth: the economics of sustainable development. Boston:Beacon Press; 1996.
[28] Davies EGR, Simonovic SP. ANEMI: a new model for integrated assessment ofglobal change. Interdiscip Environ Rev 2010;11:127e61. http://dx.doi.org/10.1504/IER.2010.037903.
[29] De Castro C. Escenarios de Energía-Economía mundiales con modelos dedin�amica de sistemas. Valladolid, Spain: University of Valladolid; 2009.
[30] De Castro Carlos, Carpintero �Oscar, Frechoso Fernando, Mediavilla Margarita,de Miguel Luis J. A Top-down approach to assess physical and ecologicallimits of biofuels. Energy Enero 2014;64:506e12. http://dx.doi.org/10.1016/j.energy.2013.10.049.
[31] De Castro C, Mediavilla M, Miguel LJ, Frechoso F. Global wind power po-tential: physical and technological limits. Energy Policy 2011;39:6677e82.http://dx.doi.org/10.1016/j.enpol.2011.06.027.
[32] De Castro C, Mediavilla M, Miguel LJ, Frechoso F. Global solar electric po-tential: a review of their technical and sustainable limits. Renew SustainEnergy Rev 2013b;28:824e35. http://dx.doi.org/10.1016/j.rser.2013.08.040.
[33] De Castro C, Miguel LJ, Mediavilla M. The role of non conventional oil in theattenuation of peak oil. Energy Policy 2009;37:1825e33. http://dx.doi.org/10.1016/j.enpol.2009.01.022.
[34] De Vries Bert JM, van Vuuren Detlef P, Hoogwijk Monique M. Renewableenergy sources: their global potential for the first-half of the 21st century ata global level: an integrated approach. Energy Policy Abril 2007;35(4):2590e610. http://dx.doi.org/10.1016/j.enpol.2006.09.002.
[35] Michael Dittmar. The end of cheap uranium. Sci. Total Environ. Setiembre2013; 461e462:792e8. http://dx.doi.org/10.1016/j.scitotenv.2013.04.035.
[36] Ehrlich PR. The limits to substitution: meta-resource depletion and a neweconomic-ecological paradigm. Ecol Econ 1989;1:9e16. http://dx.doi.org/10.1016/0921-8009(89)90021-9.
[37] European Comission. Limiting Global Climate Change to 2 degrees Cel-siusethe way ahead for 2020 and beyond. (No. COM(2007) 2 final),Communication from the Commssion. Brussels: Commission of the EuropeanCommunities; 2007.
[38] EVI IEA. Global EV Outlook. Understanding the electric vehicle landscape to2020. Electric Vehicles Initiative. International Energy Agency; 2013.
[39] EWG. Uranium resources and nuclear Energy (No. 1/2006), EWG-series.Energy Watch Group; 2006.
[40] EWG. Coal: Resources and future production (No. EWG-Paper No. 1/07); 2007.[41] EWG. Crude oil e the Supply outlook. Energy Watch Group/Ludwig-Boel-
kow-Foundation; 2008.[42] EWG. Fossil and nuclear fuels e the supply outlook (No. 2013/03/18 LBST).
Energy Watch Group; 2013.[43] FAO. The state of food insecurity in the world 2012. 2012.[44] FAOSTAT. Statistics Division of the FAO. 2014.[45] Fargione J, Hill J, Tilman D, Polasky S, Hawthorne P. Land clearing and the
[46] Fiddaman TS. Exploring policy options with a behavioral climateeeconomymodel. Syst Dyn Rev 2002;18:243e67. http://dx.doi.org/10.1002/sdr.241.
[47] Field CB, Campbell JE, Lobell DB. Biomass energy: the scale of the potentialresource. Trends Ecol Evol 2008;23:65e72. http://dx.doi.org/10.1016/j.tree.2007.12.001.
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666 665
[48] Fischedick M, Esken A, Pastowski A, Schüwer D, Supersberger N, Nitsch J,et al. Ecological, economic and structural comparison of renewable Energytechnologies (RE) with carbon capture and storage (CCS) d an integratedapproach. Wuppertal Institute for Climate, Environment and Energy;German Aerospace Center;Centre for Solar Energy and Hydrogen Research;Potsdam Institute for Climate Impact Research; 2008.
[49] Fouquet R. The slow search for solutions: lessons from historical energytransitions by sector and service. Energy Policy 2010;38:6586e96. http://dx.doi.org/10.1016/j.enpol.2010.06.029.
[50] Furtado AT, Suslick SB. Forecasting of petroleum consumption in Brazil usingthe intensity of energy technique. Energy Policy 1993;21:958e68. http://dx.doi.org/10.1016/0301-4215(93)90184-H.
[51] García D. A new world model including energy and climate change data. In:First International workshop mission earth, modeling and simulation for asustainable future; 2009. Zurich.
[52] Girod B, Wiek A, Mieg H, Hulme M. The evolution of the IPCC's emissionsscenarios. Environ Sci Policy 2009;12:103e18. http://dx.doi.org/10.1016/j.envsci.2008.12.006.
[53] Grushevenko E, Grushevenko D. Unconventional oil potential tends tochange the world oil market. Energy Sci Technol 2012;4:68e74. http://dx.doi.org/10.3968/j.est.1923847920120401.178.
[54] Guseo R. Worldwide cheap and heavy oil productions: a long-term energymodel. Energy Policy 2011;39:5572e7. http://dx.doi.org/10.1016/j.enpol.2011.04.060.
[55] Haberl H, Sprinz D, Bonazountas M, Cocco P, Desaubies Y, Henze M, et al.Correcting a fundamental error in greenhouse gas accounting related tobioenergy. Energy Policy 2012;45:18e23. http://dx.doi.org/10.1016/j.enpol.2012.02.051.
[56] Hamilton JD. Causes and consequences of the oil shock of 2007-08. Workingpaper No. 15002. National Bureau of Economic Research; 2009.
[57] Hansen J, Sato M, Kharecha P, Beerling D, Berner R, Masson-Delmotte V, et al.Target atmospheric CO2: where should humanity aim? Open Atmos Sci J2008;2:217e31. http://dx.doi.org/10.2174/1874282300802010217.
[58] Hansen J, Steffen K, Epstein P, Rohling EJ, Zachos JC, Smith P, et al. The casefor young people and nature: a path to a healthy, natural, prosperous future.2011.
[59] Harmsen JHM, Roes AL, Patel MK. The impact of copper scarcity on the ef-ficiency of 2050 global renewable energy scenarios. Energy 2013;50:62e73.http://dx.doi.org/10.1016/j.energy.2012.12.006.
[60] Heinberg R, Fridley D. The end of cheap coal. Nature 2010;468:367e9. http://dx.doi.org/10.1038/468367a.
[61] Hirsch RL. Mitigation of maximum world oil production: shortage scenarios.Energy Policy 2008;36:881e9. http://dx.doi.org/10.1016/j.enpol.2007.11.009.
[62] Holttinen H, Meibom P, Orths A, Lange B, O'Malley M, Tande JO, et al. Impactsof large amounts of wind power on design and operation of power systems,results of IEA collaboration. Wind Energy 2011;14:179e92. http://dx.doi.org/10.1002/we.410.
[63] H€o€ok M, Fantazzini D, Angelantoni A, Snowden S. Hydrocarbon liquefaction:viability as a peak oil mitigation strategy. 2013.
[64] H€o€ok M, Tang X. Depletion of fossil fuels and anthropogenic climatechangeda review. Energy Policy 2013;52:797e809. http://dx.doi.org/10.1016/j.enpol.2012.10.046.
[65] H€o€ok M, Zittel W, Schindler J, Aleklett K. Global coal production outlooksbased on a logistic model. Fuel 2010;89:3546e58. http://dx.doi.org/10.1016/j.fuel.2010.06.013.
[66] Howarth RW, Santoro R, Ingraffea A. Methane and the greenhouse-gasfootprint of natural gas from shale formations. Clim Change 2011;106:679e90. http://dx.doi.org/10.1007/s10584-011-0061-5.
[67] Hubbert MK. Nuclear energy and the fossil fuel. In: Drilling and productionpractice. American Petroleum Institute; 1956.
[69] IPCC. Climate Change 2001: synthesis report. A contribution of workinggroups I, II, and III to the third assessment report of the Integovernmentalpanel on climate change. 2001.
[72] IPCC. Mitigation of Climate Change - Contribution of Working Group III.Cambridge University Press; 2007.
[73] IPCC. Special report on renewable Energy sources and climate change miti-gation. United Kingdom and New York, NY, USA: Cambridge UniversityPress; 2011.
[74] IPCC. Impacts, adaptation and vulnerability. Contribution of working group IIto the Fifth assessment report of the Intergovernmental panel on climatechange, 2014. Cambridge University Press; 2014.
[75] IPCC SRES. Special report on emissions scenarios. 2000.[76] Jackson T. Prosperity without growth: economics for a finite planet. Reprint.
Routledge; 2009.[77] Janda K, Kristoufek L, Zilberman D. Biofuels: policies and impacts. A review.
Agric Econ e UZEI 2012;58(8):372e86.[78] Kallis G, Kerschner C, Martínez-Alier J. The economics of degrowth. Ecol Econ
2012;84:172e80. http://dx.doi.org/10.1016/j.ecolecon.2012.08.017.[79] Kerschner C. Economic de-growth vs. steady-state economy. J Clean Prod
[80] Kubiszewski I, Costanza R, Franco C, Lawn P, Talberth J, Jackson T, et al.Beyond GDP: measuring and achieving global genuine progress. Ecol Econ2013;93:57e68. http://dx.doi.org/10.1016/j.ecolecon.2013.04.019.
[81] Laherr�ere J. Oil and gas, what future?. Presented at the Groningen annualEnergy Convention, Groningen, Nederlands. 2006.
[82] Laherr�ere J. Peak oil y Seguridad Energ�etica. Presented at the SegundoSimposio ASPO Argentina Buenos Aires, Buenos Aires (Argentina). 2010.
[83] Le Page Y, Hurtt G, Thomson AM, Bond-Lamberty B, Patel P, Wise M, et al.Sensitivity of climate mitigation strategies to natural disturbances. Environ.Res. Lett. 2013;8. http://dx.doi.org/10.1088/1748-9326/8/1/015018.
[84] Lightfoot HD, Green C. Energy intensity decline implications for stabilizationof atmospheric CO2 content. No. Report No. 2001-7, October 2001. McGillCentre for Climate and Global Change Research (C2GCR); 2002.
[85] Maggio G, Cacciola G. When will oil, natural gas, and coal peak? Fuel2012;98:111e23. http://dx.doi.org/10.1016/j.fuel.2012.03.021.
[86] Manne AS, Richels RG. MERGE: an integrated assessment model for globalclimate change. In: Energy and environment; 2004. p. 175e89.
[87] Maugeri L. Oil: the next revolution. Discussion paper No. Discussion paper#2012-10. Geopolitics of Energy Project Belfer Center for Science and In-ternational Affairs; 2012.
[88] McCollum D, Bauer N, Calvin K, Kitous A, Riahi K. Fossil resource and energysecurity dynamics in conventional and carbon-constrained worlds. ClimChange 2014;123:413e26. http://dx.doi.org/10.1007/s10584-013-0939-5.
[89] MEA. Millennium Ecosystem Assessment. Ecosystems and human well-being: scenarios. Global assessment reports. Island Press; 2005.
[90] Meadows DH, Meadows DL, Randers J, Behrens III WW. The limits to growth.Universe Books; 1972.
[91] Meadows DH, Randers J, Meadows DL. Beyond the limits: confronting globalcollapse, envisioning a sustainable future. Chelsea Green Publishing Com-pany; 1993.
[92] Meadows DH, Randers J, Meadows DL. The limits to growth: the 30-yearupdate. White river Junction, Vt. Chelsea Green Publishing Company; 2004.
[93] Mediavilla M, de Castro C, Capell�an I, Javier Miguel L, Arto I, Frechoso F. Thetransition towards renewable energies: physical limits and temporal con-ditions. Energy Policy 2013;52:297e311. http://dx.doi.org/10.1016/j.enpol.2012.09.033.
[94] Mills A, Wiser R, Porter K. The cost of transmission for wind energy in theUnited States: a review of transmission planning studies. Renew SustainEnergy Rev 2012;16:1e19. http://dx.doi.org/10.1016/j.rser.2011.07.131.
[95] Mohr SH. Fossil fuel future production, world and Australia focus. Presentedat the Australian Frontiers of Science 2012: Science for a green economy,Sydney, 2e4 December 2012.
[96] Mohr SH, Evans GM. Forecasting coal production until 2100. Fuel 2009;88:2059e67. http://dx.doi.org/10.1016/j.fuel.2009.01.032.
[97] Mosconi J-J. The energy outlook in 2030 according to total. 2008.[98] Murphy DJ, Hall CAS. Year in reviewdEROI or energy return on (energy)
invested. Ann NY Acad Sci 2010;1185:102e18. http://dx.doi.org/10.1111/j.1749-6632.2009.05282.x.
[99] Murphy DJ, Hall CAS. Adjusting the economy to the new energy realities ofthe second half of the age of oil. Ecol Model 2011;223:67e71. http://dx.doi.org/10.1016/j.ecolmodel.2011.06.022.
[100] Murray J, King D. Climate policy: oil's tipping point has passed. Nature2012;481:433e5. http://dx.doi.org/10.1038/481433a.
[101] Nakicenovic N, Riahi K. Model runs with MESSAGE in the context of thefurther development of the Kyoto-Protocol, WBGUdGerman AdvisoryCouncil on Global Change. WBGU website; 2003.
[102] Nel WP, Cooper CJ. Implications of fossil fuel constraints on economic growthand global warming. Energy Policy 2009;37:166e80. http://dx.doi.org/10.1016/j.enpol.2008.08.013.
[103] Nilsson S, Schopfhauser W. The carbon-sequestration potential of a globalafforestation program. Clim Change 1995;30:267e93. http://dx.doi.org/10.1007/BF01091928.
[104] Odell PR. Why carbon fuels will dominate the 21st century's global energyeconomy. Brentwood, England: Multi-Science Pub. Co.; 2004.
[105] OECD. Looking to 2060: long-term global growth prospects. OECD EconomicsDepartment Policy Notes, no 15. Paris: OECD; 2012.
[106] Olmstead SM, Muehlenbachs LA, Shih J-S, Chu Z, Krupnick AJ. Shale gasdevelopment impacts on surfacewater quality in Pennsylvania. ProcNatl AcadSci USA 2013;110:4962e7. http://dx.doi.org/10.1073/pnas.1213871110.
[107] Osborn SG, Vengosh A, Warner NR, Jackson RB. Methane contamination ofdrinkingwateraccompanyinggas-welldrillingandhydraulic fracturing. ProcNatlAcad Sci USA 2011;108:8172e6. http://dx.doi.org/10.1073/pnas.1100682108.
[108] Pielke R, Wigley T, Green C. Dangerous assumptions. Nature 2008;452:531e2. http://dx.doi.org/10.1038/452531a.
[109] Postel SL. Entering an era of water scarcity: the challenges ahead. Ecol Appl2000;10:941e8. http://dx.doi.org/10.1890/1051-0761(2000)010[0941:EAEOWS]2.0.CO;2.
[110] Prieto PA, Hall CAS. Spain's photovoltaic revolution: the Energy return oninvestment. 2013th ed. Springer; 2013
[111] Randers J. From limits to growth to sustainable development or SD (sus-tainable development) in a SD (system dynamics) perspective. Syst Dyn Rev2000;16:213e24.
[112] Reynolds DB. The mineral economy: how prices and costs can falsely signaldecreasing scarcity. Ecol Econ 1999;31:155e66. http://dx.doi.org/10.1016/S0921-8009(99)00098-1.
I. Capell�an-P�erez et al. / Energy 77 (2014) 641e666666
[113] Rockstr€om J, Lannerstad M, Falkenmark M. Assessing the water challenge ofa new green revolution in developing countries. Proc Natl Acad Sci USA2007;104:6253e60. http://dx.doi.org/10.1073/pnas.0605739104.
[114] Rockstr€om J, Steffen W, Noone K, Persson Å, Chapin FS, Lambin EF, et al.A safe operating space for humanity. Nature 2009;461:472e5. http://dx.doi.org/10.1038/461472a.
[115] Rutledge D. Estimating long-term world coal production with logit andprobit transforms. Int J Coal Geol 2011;85:23e33. http://dx.doi.org/10.1016/j.coal.2010.10.012.
[116] Saddler H, Diesendorf M, Denniss R. Clean energy scenarios for Australia.Energy Policy 2007;35:1245e56. http://dx.doi.org/10.1016/j.enpol.2006.03.013.
[117] Schade C, Pimentel D. Population crash: prospects for famine in the twenty-first century. Environ Dev Sustain 2010;12:245e62. http://dx.doi.org/10.1007/s10668-009-9192-5.
[118] Schenk NJ, Moll HC. The use of physical indicators for industrial energydemand scenarios. Ecol Econ 2007;63:521e35. http://dx.doi.org/10.1016/j.ecolecon.2006.12.008.
[119] Schneider M, Froggatt A, Hazemann J. The world nuclear Industry statusreport 2012. 2012.
[120] Scott V, Gilfillan S, Markusson N, Chalmers H, Haszeldine RS. Last chance forcarbon capture and storage. Nat Clim Change 2013;3:105e11. http://dx.doi.org/10.1038/nclimate1695.
[121] Searchinger T, Heimlich R, Houghton RA, Dong F, Elobeid A, Fabiosa J, et al.Use of U.S. Croplands for biofuels increases Greenhouse gases throughemissions from land-use change. Science 2008;319:1238e40. http://dx.doi.org/10.1126/science.1151861.
[122] Simon JL. The ultimate resource 2. Princeton, NJ: Princeton University Press;1996.
[123] Skrebowski C. The Oil Crunch: a wake-up call for the UK economy. UK In-dustry Taskforce on Peak Oil & Energy Security (ITPOES); 2010.
[124] Smil V. Energy transitions: history, requirements, Prospects. Praeger; 2010.[126] S€oderbergh B, Robelius F, Aleklett K. A crash programme scenario for the
Canadian oil sands industry. Energy Policy 2007;35:1931e47. http://dx.doi.org/10.1016/j.enpol.2006.06.007.
[128] Sorrell S, Speirs J, Bentley R, Brandt A, Miller R. Global oil depletion. Anassessment of the evidence for a near-term peak in global oil production. UKEnergy Research Centre; 2009.
[129] Sterman JD. System dynamics modeling. Calif Manag Rev 2001;43:8e25.[130] Stern DI. Limits to substitution and irreversibility in production and
consumption: a neoclassical interpretation of ecological economics. EcolEcon 1997;21:197e215. http://dx.doi.org/10.1016/S0921-8009(96)00103-6.
[131] Teske S, Pregger T, Simon S, Naegler T, Graus W, Lins C. Energy [R]evolution2010da sustainable world energy outlook. Energy Efficiency 2011;4:409e33. http://dx.doi.org/10.1007/s12053-010-9098-y.
[132] Thielemann T. Why there will be No peak coal in the foreseeable future.Open J Geol 2012;02:57e64. http://dx.doi.org/10.4236/ojg.2012.22006.
[133] Trainer T. A critique of Jacobson and Delucchi's proposals for a worldrenewable energy supply. Energy Policy 2012;44:476e81. http://dx.doi.org/10.1016/j.enpol.2011.09.037.
[134] Turner GM. A comparison of the limits to growth with 30 years of reality.Glob Environ Change 2008;18:397e411. http://dx.doi.org/10.1016/j.gloenvcha.2008.05.001.
[135] Turner GM. On the Cusp of global collapse? updated comparison of the limitsto growth with historical data. GAIA e Ecol Perspect Sci Soc 2012;21:116e24.
[136] Tverberg GE. Oil supply limits and the continuing financial crisis. Energy2012;37:27e34. http://dx.doi.org/10.1016/j.energy.2011.05.049.
[137] UNEP. The GEO-3 scenarios, 2002-2032: quantification and analysis ofenvironmental impacts, UNEP/DEWA/RS. Nairobi, Kenya: Bilthoven,Netherlands: UNEP; RIVM; 2004.
[138] UNEP. Global environment outlook: environment for development, GEO 4.Nairobi, Kenya; London: United Nations Environment Programme; Statio-nery Office; 2007.
[139] UNEP. Towards sustainable production and use of resources: assessing bio-fuels. 2009.
[140] UNEP. Global environment outlook GEO 5: environment for the future wewant. Nairobi, Kenya: United Nations Environment Program; 2012.
[141] US EIA db. International Energy Statistics (Database). US Energy InformationAdministration; 2014. http://www.eia.gov/cfapps/ipdbproject/IEDIndex3.cfm.
[142] Valero A, Valero A. Physical geonomics: combining the exergy and Hubbertpeak analysis for predicting mineral resources depletion. Resour ConservRecycl 2010;54:1074e83. http://dx.doi.org/10.1016/j.resconrec.2010.02.010.
[143] Van den Bergh JCJM. The GDP paradox. J Econ Psychol 2009;30:117e35.http://dx.doi.org/10.1016/j.joep.2008.12.001.
[144] Van Vuuren DP, Kok MTJ, Girod B, Lucas PL, de Vries B. Scenarios in GlobalEnvironmental Assessments: key characteristics and lessons for future use.Glob Environ Change 2012;22:884e95. http://dx.doi.org/10.1016/j.gloenvcha.2012.06.001.
[145] Ward JD, Mohr SH, Myers BR, Nel WP. High estimates of supply constrainedemissions scenarios for long-term climate risk assessment. Energy Policy2012;51:598e604. http://dx.doi.org/10.1016/j.enpol.2012.09.003.
[146] WBGU. Future bioenergy and sustainable land use. 2008.[147] WEC. Survey of energy resources. World Energy Council; 2010.[148] WEO. World energy outlook 2010. OECD/IEA; 2010.[149] WEO. World energy outlook 2011. Paris: IEA, international Energy Agency:
OECD; 2011.[150] WEO. World energy outlook 2012. OECD/IEA; 2012.[151] Zerta M, Schmidt PR, Stiller C, Landinger H. Alternative World Energy
Outlook (AWEO) and the role of hydrogen in a changing energy landscape.Int J Hydrogen Energy 2008;33:3021e5. http://dx.doi.org/10.1016/j.ijhydene.2008.01.044.
[152] Zittel W. Feasible futures for the common Good. energy transition. Paths in aPeriod of Increasing Resource Scarcities. 2012.
[153] World Bank database. World Bank database. http://data.worldbank.org/;2014.
[154] European Commission. REPORT FROM THE COMMISSION on indirect land-use change related to biofuels and bioliquids (No. COM(2010) 811 final);2010.
![Satelites[1]. Articulo Cientifico](https://static.documents.pub/doc/80x56/5571f9a34979599169900e74/satelites1-articulo-cientifico.jpg)
