Standard, Point of Use, and Extended Energy Return on Energy Invested (EROI) from Comprehensive Material Requirements of Present Global Wind, Solar, and Hydro Power Technologies

energies Article Standard, Point of Use, and Extended Energy Return on Energy Invested (EROI) from Comprehensive Material Requirements of Present Global Wind, Solar, and Hydro Power Technologies Carlos de Castro 1,2,* and Iñigo Capellán-Pérez 1 1Research Group on Energy, Economy and System Dynamics, Escuela de Ingenierías Industriales, University of Valladolid, Paseo del Cauce s/n, 47011 Valladolid, Spain; [email protected] 2Department of Applied Physics, Escuela de Arquitectura, University of Valladolid, Av Salamanca, 18, 47014 Valladolid, Spain *Correspondence: ccastr[email protected] Received: 24 April 2020; Accepted: 6 June 2020; Published: 12 June 2020   Abstract: Whether renewable energy sources (RES) will provide sufficient energy surplus to entirely power complex modern societies is under discussion. We contribute to this debate by estimating the current global average energy return on energy invested (EROI) for the five RES technologies with the highest potential of electricity generation from the comprehensive and internally consistent estimations of their material requirements at three distinct energy system boundaries: standard farm-gate (EROI st ), final at consumer point-of-use (EROI final ), and extended (including indirect investments, EROI ext ). EROI st levels found fall within the respective literature ranges. Expanding the boundaries closer to the system level, we find that only large hydroelectricity would currently have a high EROI ext ~ 6.5:1, while the rest of variable RES would be below 3:1: onshore wind (2.9:1), offshore wind (2.3:1), solar Photovoltaic (PV) (1.8:1), and solar Concentrated Solar Power (CSP) (<1:1). These results indicate that, very likely, the global average EROI ext levels of variable RES are currently below those of fossil fuel-fired electricity. It remains unknown if technological improvements will be able to compensate for factors, which will become increasingly important as the variable RES scale-up. Hence, without dynamically accounting for the evolution of the EROI of the system, the viability of sustainable energy systems cannot be ensured, especially for modern societies pursuing continuous economic growth. Keywords: EROI; transition to renewables; wind; solar; large hydropower 1. Introduction The solution to the coupled problems of climate change, pollution, and the depletion of fossil fuels (FF) needs a rapid transition, in historical terms, from the current fossil-based energy system to one whose main energy sources will be renewable (renewable energy sources, RES). In the ongoing debate aboutthetimingandthe featuresofthistransition, environmental, economic, political, and technological issues are pointed out [ 1 , 2 ]. With relation to the latter, the discussion has mainly addressed three interrelated aspects: 1. The requirements and future availability of minerals to build the infrastructures to harness, transform, transport, and store energy [3–6]. 2. The maximum technological, economic and sustainable energy flows available (potentials) for each energy technology/resource [7–12]. Energies 2020,13, 3036; doi:10.3390/en13123036 www.mdpi.com/journal/energies Energies 2020,13, 3036 2 of 42 3. The energy surplus generated by the different energy sources available to society for discretionary uses (from growing food to family support, as well as education, health, and cultural development [ 13 ]). The most popular indicator to measure the ratio of energy surplus with relation to the required energy investments is the energy return on energy invested (EROI), i.e., the ratio of the energy delivered and the energy consumed to deliver that energy in a given time, with very abundant literature focusing both at technology and system level at different geographical scales (see e.g., [3,14–23]). Although much work has been carried out to estimate the EROI of individual RES technologies, important methodological discrepancies exist depending on the applied definition, the system design and location, the functional units, or the boundaries of the analysis (i.e., mine-mouth vs. final use or individual energy technology vs. energy system) [ 16 , 21 , 24 – 31 ]. Different methods have been developed in the literature [ 3 , 16 , 29 , 32 , 33 ]; however, the inter-comparison of the EROIs obtained by these diverse methods is far from being straightforward (and often even not possible in a consistent way). Hence, although the EROI is generally acknowledged as a relevant factor for robustly designing the energy transition, and many multidisciplinary studies and meta-analyses have been performed with the aim of synthetizing and standardizing the state-of-the-art of the field [ 14 , 19 , 20 , 25 , 34 – 36 ], this dispersion of methodologies, definitions, and results hinders the potential to use the EROI concept to adequately inform policy makers and society as a whole. One of the open debates of greatest importance is whether the RES with higher potential have a sufficient EROI to maintain the energy “metabolism” of a large and complex civilization, such as the industrial-one. In fact, a favorable EROI over the long-term has been associated in fields such as biology or anthropology as a key driver of increasing complexity and evolution for plants, animals and humans [ 37 – 40 ]. However, the question on the minimum EROI for human societies remains arduous and elusive in the literature [ 13 , 38 , 41 , 42 ]. It has been indirectly approached in the literature through the comparison with FF under the logic that RES were found to have an EROI level at least as high as the one characterizing FF, then it could be assured that RES could also fuel complex modern societies. On the one hand, those RES with a higher potential (i.e., wind, solar) have been generally found to have lower EROI standard (EROI st ) than FF, especially when incorporating the energy costs of dealing with their variability [ 19 ]. However, some recent works have found that at the electricity generated level, the EROI of FF declines substantially, making that the EROI of RES could be, nowadays, in fact, better than the one generated by FF [ 16 , 33 , 43 ]. However, the comparison is not straightforward. Brockway et al. [ 16 ] use an input–output extended approach for estimating the yearly EROI of all aggregated FF at a global level without accounting for capital investments, which prevented the application of the same method for RES (hence, performing the comparison through literature review). Raugei [ 33 ] and Raugei and Leccisi [ 43 ] refer to some FF particular cases at a national level, taking, as a timeframe, the full lifetime of the power plants at the point of use level. It should also be taken into account that static EROI computations (over the lifetime) are not able to capture dynamic issues inherent to the dynamic nature of the energy transition, such as the requirement of up-front investments for the installation of new RES power plants [ 3 ] or the decreasing historical trend in the EROI of FF [16,19,44,45]. In this work, we contribute to this ongoing debate by estimating the current global EROI over the lifetime for the five RES energy technologies assessed to have the highest theoretical potential for electricity generation (solar Photovoltaic (PV) and Concentrated Solar Power (CSP), wind onshore, wind offshore and large hydroelectricity) [ 11 , 46 ]. We compute the EROI at standard or primary and point-of-use or final, and estimate the EROI at extended boundaries (full-system level). This approach allows us to identify those steps where more energy investments are required, as well as a consistent inter-comparison between the EROI levels of different technologies. Two main novelties are applied with relation to the current state-of-the-art. First, an extensive and comprehensive literature review is performed in order to collate internally consistent data about the material requirements of each technology. In the cases where published data for an element/phase Energies 2020,13, 3036 3 of 42 of the manufacture/installation of the technology was not found, the material requirements have conservatively been estimated from available data from other technologies (instead of being assumed 0 as most commonly performed in the literature). Second, we focus on performance parameters of real systems, globally, which contrasts with the widespread approach in the literature applying theoretical assumptions for modeling RES systems and/or focus on particular plants or countries (a method that frequently leads to “cherry picking”), which has been shown to overestimate the performance of real technologies and systems [25,47–54]. This work documents in detail the assumptions and data used to estimate the material and energy investments to install, operate, and maintain new RES infrastructures used in the MEDEAS modeling framework [ 3 , 55 , 56 ]. In this framework, the computation of the “static” (over the lifetime) EROI at three levels (standard, point of use, and extended) of the different energy technologies is an intermediate step to consistently integrate the mineral and energy requirements of RES in the wider modeling framework, able to represent the dynamic nature of the energy transition, as shown in a previous analysis on dynamic EROI based at the standard level [ 3 ]. This work reports updated and original results based on a more detailed lifecycle analysis, which have still not integrated in the MEDEAS models. Section 2overviews and compares the different definitions of EROI in the literature and Section 3 covers the methodology applied to estimate the EROI levels in this work. Results are presented and discussed in Sections 4and 5, respectively, and Section 6concludes. 2. Overview of EROI Definitions and Their Implications for Society A diversity of different methods to estimate the EROI of a given technology, energy resource, or full energy system exists in the literature, according to different criteria to select the system boundaries and interpret which is the energy “usable” by the society, different definitions, etc. [3,16,29,32,33]. In this work, we estimate the EROI at the three boundaries most common in the literature: the EROI standard or primary (EROI st , Equation (1)), EROI at the point of use or final (EROI final , Equation (2)) and EROI extended (EROI ext , Equation (3)). The EROI st includes the on-site and offsite (i.e., energy needed to manufacture the devices and systems used later on site) energy requirements to get the energy (e.g., build, operate, and maintain a power plant). As the energy investments are usually computed in “primary” energy terms and the energy delivered is computed in final terms, a “quality” correction factor (g) is typically used to compare different technologies and sources [Note that the authors have elsewhere (cf. [ 3 ]) defined a different EROI st (between the classical EROI st used here and the EROI final ) applied to the system level, including the energy investments associated to the management of renewable energy sources (RES) variability, as well as considering that the real useful energy for the society is the one reaching the consumers rather than the one delivered at the mouth-gate of the power plants]. The EROI final includes the energy costs to get and deliver the energy carrier to the point of use of society (e.g., refining, transportation, etc.). The EROI ext includes the energy required to get and deliver a unit of energy (EnU direct ), as well as the indirect energy (EnU indirect , or supply chain energy investment in input-output (IO) terminology) required to produce the machinery and devices used to build, operate, and maintain a power plant or a transportation facility (tank truck, pipeline, etc.), as well as the energy required for exploration, investment, communication, labor, etc. Hence, the EROI ext is the most significant EROI type, given that the discretionary uses of energy are the relevant ones at societal level. EROIst =Energy delivered by the plant or energy source Energy used to deliver the energy by the energy system =Efinal, out EnUprimary ·g(1) EROIfinal =Final energy delivered to the final consumer Direct final energy used to deliver the energy =Efinal, out EnUdirect (2) EROIext =Final energy delivered to the final consumer Direct +indirect final energy used to deliver the energy =Efinal,out EnUdirect +EnUindirect (3) Energies 2020,13, 3036 4 of 42 Figure 1shows graphically the three aforementioned definitions of EROI (standard, point-of-use or final and extended) based on a flow chart. Energies 2020, 13, x FOR PEER REVIEW 4 of 43 Figure 1 shows graphically the three aforementioned definitions of EROI (standard, point-of-use or final and extended) based on a flow chart. Figure 1. Energy flux from primary material sources of energy {0} to discretionary uses of energy. EROIst = ({0}/g)/{2} (EROI standard, being “g“ a correction factor to convert “primary” to final). EROIfinal = {1}/{2′} (EROI final, where {2′} includes the direct use of energy to bring it to the user) and EROIext = {1}/({2’+3}) (EROI extended, all the terms at final and user phase). Note that the relevant energy for a society is the blue box (discretionary uses of energy), but primary sources of energy, and the other boxes, are relevant for environment issues because they are always taken/used in the biosphere (see reference [3] for more details). Figure 2 shows the implications for an energy system that maintains the same discretionary energy flow when the EROIfinal decreases from 9:1 to 3:1, even with less relative losses (greater factor {1}/{0}). To achieve the same level of discretionary uses of energy with a lower EROI of the system, the economic sectors related to energy production and the primary energy sources to be extracted and processed increase substantially (final energy processed by society increases 2.33 times in the example depicted in Figure 2). Few works have, to date, dealt with the intricate issue of the minimum EROI to sustain modern complex societies. Different works, applying different methodologies [13,41,42], have suggested that a minimum EROIst of the system > 10–15:1 is required to sustain advanced industrial societies. However, these estimations are approximates being subject to important limitations, such as the reliance on prices of energies and, in some cases, are just educated guesses without calculations backing them. Hall et al. [38] went a step forward, finding that a minimum of EROIst = 3:1 for oil and liquid biofuels is required to reach the final consumer in the USA. Other studies, such as Brandt [57], have approached the topic from a more theoretical point of view. Moreover, these estimates do not reach the EROIext, which is the relevant boundary at system level. In fact, although an energy technology with EROIext >1:1 is a net source of energy for the society, a complex modern society with an EROIext very close to 1:1 would still not be viable due to at least four reasons: Discretionary uses of energy {1-2-3} Final energy for society {1} energyNet to society {1-2} Primary sources{0} {2} Direct use of energy for energy {3} Indirect use of energy for energy Losses {0-1} Figure 1. Energy flux from primary material sources of energy {0} to discretionary uses of energy. EROIst =({0}/g)/{2} (EROI standard, being “g“ a correction factor to convert “primary” to final). EROI final ={1}/{2 0 } (EROI final, where {2 0 } includes the direct use of energy to bring it to the user) and EROI ext ={1}/({2’+3}) (EROI extended, all the terms at final and user phase). Note that the relevant energy for a society is the blue box (discretionary uses of energy), but primary sources of energy, and the other boxes, are relevant for environment issues because they are always taken/used in the biosphere (see reference [3] for more details). Figure 2shows the implications for an energy system that maintains the same discretionary energy flow when the EROI final decreases from 9:1 to 3:1, even with less relative losses (greater factor {1}/{0}). To achieve the same level of discretionary uses of energy with a lower EROI of the system, the economic sectors related to energy production and the primary energy sources to be extracted and processed increase substantially (final energy processed by society increases 2.33 times in the example depicted in Figure 2). Few works have, to date, dealt with the intricate issue of the minimum EROI to sustain modern complex societies. Different works, applying different methodologies [ 13 , 41 , 42 ], have suggested that a minimum EROI st of the system >10–15:1 is required to sustain advanced industrial societies. However, these estimations are approximates being subject to important limitations, such as the reliance on prices of energies and, in some cases, are just educated guesses without calculations backing them. Hall et al. [ 38 ] went a step forward, finding that a minimum of EROI st =3:1 for oil and liquid biofuels is required to reach the final consumer in the USA. Other studies, such as Brandt [ 57 ], have approached the topic from a more theoretical point of view. Moreover, these estimates do not reach the EROI ext , which is the relevant boundary at system level. In fact, although an energy technology with EROI ext >1:1 is a net source of energy for the society, a complex modern society with an EROI ext very close to 1:1 would still not be viable due to at least four reasons: Energies 2020,13, 3036 5 of 42 Energies 2020, 13, x FOR PEER REVIEW 5 of 43 Figure 2. A comparison between two hypothetical societies with the same energy flow available for discretionary uses {1-2′-3} (blue box) but with different EROIfinal and relative losses ({0-1}/{1}). The size of vertical arrows is at scale. (a) EROIfinal = 9:1, {1}/{0} = 0.7, assuming indirect energy use equal to direct one (see Section 3); (b) EROIfinal= 3:1, {1}/{0} = 0.8, indirect energy used equal to direct one ({2′} = {3}); (c) shows the resulting share of net power vs. gross power in the net energy cliff.  A low EROIext means that the energy metabolism ({0}, {1}, {2}, and {3}) represents a substantial part of the economy. Since each energy flux requires capital and workers, in that case the society allocates most parts of its capital and workers to the energy system, and not the discretionary uses. Therefore, a much lower diversity of jobs and enterprises can be attained in these conditions and the result is a much simpler society (as e.g., in pre-industrial times with most workers in the primary sector and very few in the tertiary sectors). In fact, there is an “EROI minimum”, below which a complex society tends to disappear or evolve towards simpler organizational forms [38,39,58]. In fact, when the EROI approaches 1:1, the capacity installed (and, hence, the primary energy required) tends to infinity if the same level of net energy is to be maintained (see Equations (9)–(11) in Capellán-Pérez et al. [59]).  The environmental impact depends on factors, such as the type of the energy resource (e.g., pollution and climate change caused by FFs’ combustion) and the size of the energy system (e.g., mining impacts, material residues, land-use, co-optation of fluxes of the biosphere—wind, biomass, etc.). Hence, for the same discretionary uses of energy, a lower EROI means larger environmental impacts and the need to divert a larger share of the final energy from discretionary uses to “defensive” costs [60]. Final energy for society {1} Net energy to society {1-2} Primary sources {0} a) “High” EROI EROIfinal = 9:1 {1}/{0} = 0.7 {3} = {2’} Losses {0-1} Discretionary uses of energy {1-2-3} {2} Direct use of energy for energy {3} Indirect use of energy for energy Net energy to society {1-2} Losses {0-1} Direct use of energy for energy Final energy for society {1} Primary sources {0} b) “Low” EROI EROIfinal = 3:1 {1}/{0} = 0.8 {3} = {2’} Discretionary uses of energy {1-2-3} {2} Indirect use of energy for energy {3} c) The net energy cliff EROI final = 9:1 EROI final = 3:1 0% 20% 40% 60% 80% 100% 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 Net energy (% of final energy) EROI final Figure 2. A comparison between two hypothetical societies with the same energy flow available for discretionary uses {1-2 0 -3} (blue box) but with different EROI final and relative losses ({0-1}/{1}). The size of vertical arrows is at scale. ( a ) EROI final =9:1, {1}/{0} =0.7, assuming indirect energy use equal to direct one (see Section 3); ( b ) EROI final =3:1, {1}/{0} =0.8, indirect energy used equal to direct one ({20}={3}); (c) shows the resulting share of net power vs. gross power in the net energy cliff. • A low EROI ext means that the energy metabolism ({0}, {1}, {2}, and {3}) represents a substantial part of the economy. Since each energy flux requires capital and workers, in that case the society allocates most parts of its capital and workers to the energy system, and not the discretionary uses. Therefore, a much lower diversity of jobs and enterprises can be attained in these conditions and the result is a much simpler society (as e.g., in pre-industrial times with most workers in the primary sector and very few in the tertiary sectors). In fact, there is an “EROI minimum”, below which a complex society tends to disappear or evolve towards simpler organizational forms [ 38 , 39 , 58 ]. In fact, when the EROI approaches 1:1, the capacity installed (and, hence, the primary energy required) tends to infinity if the same level of net energy is to be maintained (see Equations (9)–(11) in Capellán-Pérez et al. [59]). • The environmental impact depends on factors, such as the type of the energy resource (e.g., pollution and climate change caused by FFs’ combustion) and the size of the energy system (e.g., mining impacts, material residues, land-use, co-optation of fluxes of the biosphere—wind, biomass, etc.). Hence, for the same discretionary uses of energy, a lower EROI means larger Energies 2020,13, 3036 6 of 42 environmental impacts and the need to divert a larger share of the final energy from discretionary uses to “defensive” costs [60]. • Societies require a “security buffer” (i.e., buy “insurance”), to be able to overcome unexpected events, such as accidents or natural disasters (e.g., earthquakes). • Human inequality makes the metabolic system less efficient (in the real world, part of the discretionary energy uses will always be metabolically “useless”, such as luxuries of rich or corrupted people), so again, the supply of “useful” discretionary uses (food, domestic, education, etc.) requires an EROIext >1. • There is a critical additional reason in the context of the energy transition given that it will require the temporary fast growth of RES sources and the dismantling of the FF they replace. In this situation, the EROI of the full system will temporarily be well below the weighted average of the static EROI of the technologies and their supporting systems (e.g., grids, storage, etc.), as shown in [ 3 ]. Noteworthy, in a society where population and energy per capita are growing, this phenomenon of “energy trap” will be aggravated. It has also to be highlighted that a minimum EROI to maintain a sustainable modern complex society cannot be established, precisely given that the reduced availability of discretionary energy as intermediary operations become less efficient is a gradual non-linear process with increasing and cascade consequences over time. Hence, it can also be useful to think about ranges and increasing levels of risk (see, e.g., discussions in Brandt [ 57 ] and around Figure 9 in [ 3 ]). This minimum EROI depend also ultimately on social decisions, depending on how each society decides to allocate the available resources in case of decreasing EROI between investments and consumption taking into account also social inequalities. See the Discussion section in this paper for our views on how to approach this issue in further research. 3. Methodology This section documents the estimation of the current global EROI over the lifetime at three distinct boundaries (standard or primary, point-of-use or final and extended) for the five RES energy technologies assessed to have the higher theoretical potential for electricity generation (solar PV and CSP, wind onshore, wind offshore, and large hydroelectricity). This section starts with the selection of the representative technologies (Section 3.1), follows with the definitions of EROI used at different boundaries (Section 3.2) and the material requirements to build, operate, and dismantle each RES technology power plant (Section 3.3). The details for the estimation of the EROI of each RES technology studied are collated in Appendices Aand B. 3.1. Selection of Representative Technologies In general, a “representative” technology is selected for each alternative technology on the basis of their current and the near future expected performance. Table 1shows the selected representative technologies as well as the main references considered in this work for their material intensities. For the case of solar PV, a weighted average is computed for some minerals taking into account the current share of PV sub-technologies. Supplementary Material collates the material intensity for each technology and gives references and detailed comments. Current global average mineral recycling rates correspond to the share of recycled content (RC) in the fabricated metal from [ 61 ] (see Supplementary Material). Material losses in the gate-to-gate phase are conservatively estimated (see also Appendix C about conservative assumptions) (being recycled for other uses or not). Energies 2020,13, 3036 7 of 42 Table 1. Representative alternative technologies and main references considered for their material intensities. See Supplementary Material for details. Alternative Technology Representative Technology Main References for Material Intensities Solar Concentrated Solar Power (CSP) CSP (parabolic trough collector) with molten-salt storage without back-up: most efficient and used technology [25]. Back-up option is not considered since it is usually powered by non-renewable fuels such as natural gas. [25,62] Solar Photovoltaic (PV) Fixed-tilt silicon PV: better performance in terms of Energy inputs and EROI [21] and subject to less mineral availability constraints [63]. A weighted average is assumed taking into account the current share of thin-film technologies in global PV mix. [21,25,62,64] Wind onshore 2 MW onshore wind turbines, which is over the current global average wind onshore installed capacity per turbine [65]. [62,66] Wind offshore 3.6 MW offshore wind turbines taking as reference the current average size in Europe [67]. [62,66,68,69] Large hydroelectricity (>10 MW) Large power plant used by [ 70 ] (96 MW) (the average capacity for a large hydroelectricity power plant is ~100 MW [71]. [70] 3.2. EROI Computation as a Function of the System Boundaries 3.2.1. EROI Expressions Equations (4)–(6) show the gross standard, point-of-use or final and extended EROI expressions over the lifetime of the power system (plant or storage facility): EROIst =Eout,pp PEnUst direct = =Capacity power·CF·L·(1−OL) gst·(EnUNew cap+EnUO&M+EnUDecom+hj·EnUTra+SC (4) EROIfinal =Eout,consumer PEnUfinal direct = =Capacity power·CF·L·(1−OL)·(1−TDL) gfinal·(EnUNew cap+EnUO&M+EnUDecom+EnUG&S+hj·EnUTra+SC·(1+TDL) (5) EROIext =Eout,consumer PEnUfinal direct +EnUindirect (6) where: Eout,pp: electricity produced by the power plant. Eout,consumer: electricity delivered to the consumer. Capacity power: nominal capacity power considered: 1 MW. CF: capacity factor (dimensionless), the ratio of the actual electrical energy output over a given period of time to the maximum possible electrical energy output over the same period. L: operational lifetime of the power system in seconds. OL: operational electricity losses for each technology are estimated as a share of the electricity output of the power plant at the inverter or transformer meter, including local distribution grid losses to the transmission grid, operational shutdown losses, and expected degradation of the power plant. Double accounting with real CF is avoided by estimating the net future expected degradation, taking into consideration that most RES power plants are very young. Energies 2020,13, 3036 8 of 42 TDL: the transmission and distribution losses in the grid as a share of the electricity output of the power plant. They represent the losses from the transformer or inverter of the power plant to the final consumer. TDL =0 for EROI st and TDL =0.092 for EROI final and EROI ext (based on the ratio of power plant net production of electricity at global level and the total electricity consumed in 2015. The electricity final consumption in 2015 was 1737 Mtoe and the net electricity production of the power plants was 1913 Mtoe (2091 Mtoe gross production—178 Mtoe self-consumption of electricity) [ 72 ]. Hence, TDL =1 − 1737/1913 =0.092. Final consumption to gross production losses were thus: 1−1737/2091 =0.169; therefore, the electrical global mix in 2015 had OL =1 − (1 − 0.169)/(1 − 0.092) =0.085). SC: electricity self-consumption of the facility from the grid for auxiliary devices (lights, pumps, motors, etc.). For EROI final , the TDL from the grid to the facility are considered equal to the TDL from the facility to the consumer (here, the consumer is the power plant). EnU New cap : Final Energy Used (Joules) for the construction phase of the new installed capacity (cradle to end gate). We further divide the EnU New cap in three components (CtoG, GtoG and GtoE) to adapt to the data availability of material requirements in the construction phase (see Section 3.2.2 and Appendix A): EnUNew cap CtoG is the embodied energy in the Raw/resource to the Suppliers chain (cradle to gate in life-cycle analysis (LCA) terminology), EnUNew cap GtoG is the embodied energy from the suppliers to the finished products of the power plant or the Manufacturing phase (gate to gate phase), and EnUNew cap GtoE is the embodied energy in the Erection of the power plant. EnU O&M : final energy used (Joules) in the operation and maintenance (O&M) of the installed capacity in their lifetime. EnU Decom : final energy used (Joules) for decommissioning those infrastructures that have ended their lifetime. A 10% of the EnU New cap is assumed for all technologies following [ 73 ] due to the lack of relevant global data. Other studies use a 50 to 100% of the site preparation energy costs (see Kis et al. [74]), which deliver similar results than the 10% of the EnUNew cap used here. EnU G&S : final energy used (Joules) in grids, storage and other related infrastructures necessary to transport and manage the electricity to the point of use. For EROI st , EnU G&S =0. For the sake of simplicity, we assign the O&M material costs of grids associated to the current mix as if they were independent of the technology (see Appendix B). However, due to the lower CF of variable RES relative to the current global mix and their variability, RES will likely require more new grids and other adaptation costs per MW as they scale up. Moreover, the storage per MW requirements (measured through their ESOI) are difficult to assign to a specific technology and have more sense in a dynamic transition scenario as done in [3]. EnU indirect : relative indirect EnU to direct EnU. We use 100%, i.e., EnU direct =EnU indirect based on the comparison with meta-analysis that compare LCA versus Input–Output (IO) analysis for wind and solar (see Section 3.2.3). EnU Tra : final energy used for the transport of the materials: diesel and fuel oil (see Appendix B). Diesel and fuel oil are converted to primary energy with their respective factor h j estimated as the ratio between the well-to-wheels with the train-to-wheels (1.19 and 1.09 [ 75 ], respectively, see excel table in the supplementary material for references and details). See next section for the method to assure consistency between primary energy inputs in the EnU computation and final energy outputs (g factor) for the three EROI indicators. Figure 3shows the conceptual representation of the energy inputs and output for power plants for variable renewable electricity generation and storage facilities, considering the different phases at different plant and system boundaries. Energies 2020,13, 3036 9 of 42 Energies 2020, 13, x FOR PEER REVIEW 9 of 43 Figure 3 shows the conceptual representation of the energy inputs and output for power plants for variable renewable electricity generation and storage facilities, considering the different phases at different plant and system boundaries. Energy flow Electricity output · (1 - OL) · (1-TDL) Time Construction time EnUDecom EnUO&M + EnUTra Self-consumption Lifetime Decommision time EnUG&S + EnUTra EnU    EnU   EnU    EnU    EnU    EnU   EnUTra EnUTra Figure 3. Conceptual representation of the energy inputs and output for power plants for variable renewable electricity generation and storage facilities. Black boxes represent energy investments considered at the EROI st phase, grey boxes the additional features incorporated into the EROI final , and the white boxes the indirect investments to compute the EROIext. Adapted from [3], not to scale. Energies 2020,13, 3036 16 of 42 international trade to net import energy (and carbon) intensive products from other countries, such as the BRIICS (Brazil, Russia, India, Indonesia, China) [107]. Energies 2020, 13, x FOR PEER REVIEW 17 of 43 Figure 5. EROI levels at standard (EROIst), final (EROIfinal), and extended (EROIext) stages of PV solar on land for the top 10 countries with more installed capacity, the global average, and Finland and Cyprus as holders of the minimum and maximum irradiance at European Union (EU) level. The current weighted global average for solar irradiance where power plants are installed is found to be ~160 W/m2. 4.3. Comparison of the EROIst of RES Technologies with the Literature As discussed in Section 2, the comparison of the EROI values obtained in the present study with other studies is tricky due to the methodological discrepancies existing in the literature and the Finland; 2.8 UK; 4.4 Germany; 5.1 France; 6.5 Japan; 7.5 World; 7.7 USA; 8.0 Italy; 7.7 Spain; 8.7 China; 9.1 India; 10.2 Australia; 11.4 Cyprus; 12.5 Finland; 1.3 UK; 2.0 Germany; 2.3 France; 2.9 Japan; 3.4 World; 3.5 USA; 3.6 Italy; 3.5 Spain; 3.9 China; 4.1 India; 4.7 Australia; 5.2 Cyprus; 5.7 Finland; 0.6 UK; 1.0 Germany; 1.2 France; 1.5 Japan; 1.7 World; 1.8 USA; 1.8 Italy; 1.8 Spain; 2.0 China; 2.1 India; 2.3 Australia; 2.6 Cyprus; 2.8 0 1 2 3 4 5 6 7 8 9 10 11 12 13 75 125 175 225 275 EROI solar PV Average irradiance (W/m²) EROIst EROIfinal EROIext Figure 5. EROI levels at standard (EROI st ), final (EROI final ), and extended (EROI ext ) stages of PV solar on land for the top 10 countries with more installed capacity, the global average, and Finland and Cyprus as holders of the minimum and maximum irradiance at European Union (EU) level. The current weighted global average for solar irradiance where power plants are installed is found to be ~160 W/m 2 . Energies 2020,13, 3036 17 of 42 4.3. Comparison of the EROIst of RES Technologies with the Literature As discussed in Section 2, the comparison of the EROI values obtained in the present study with other studies is tricky due to the methodological discrepancies existing in the literature and the uncertainty associated to the data collection (see Appendixes C–Efor more details). Still, we report the comparison of our obtained results for EROI st with the literature range found in the review papers and meta-analyses to show how far or near we are of other published studies. We also compare with the EROI st calculations from Dupont and Kis et al., [ 74 , 94 , 103 ] since this team follows, to the knowledge of the authors, a methodology that is the closest to the one developed here; therefore, the direct comparison of results is the most meaningful. However, for a direct comparison a harmonization process should be endeavored with relation to the assumptions and performance factors (CF, g factor, etc.). Table 4shows that our EROI st results are within but on the lower bound of the literature for the RES technologies excepting for CSP [ 14 , 20 , 34 , 74 , 94 , 103 , 108 , 109 ]. Two main reasons explain this: (1) few studies use so many materials for the different construction, O&M, and dismantling phases, and (2) it is customary in the field to use favorable performance factors, which overestimate the average of real systems (see discussion in [ 25 ]). For CSP, our results are even below the low bound of the published ranges because, likely due to their relative low capacity installed, most studies rely excessively in theoretical grounds (as was demonstrated in [ 25 ]) and not in performance factors from real power plants. In any case, as argued in Section 2, we recall that the standard boundary for EROI is not sufficiently relevant to analyze the implications of the energy transition. Table 4. Comparison of the EROIst obtained in this work with the literature. “n” represents the number of the studies reviewed in meta-analysis studies. EROIst levels are not harmonized. Technologies EROIst This Work EROIst Literature Range Reference Literature Range Meta-Analysis and Individual Studies Large hydro 28.4 10–105 Dale [108]; n =16 11.2–267 Schoenberg and Hall [109]; n =7 5.9–49.6 (24.7) Kis et al. [74] (min-max)(base) Wind onshore 13.2 12.5–66.7 Carbajales-Dale [34] (n =42; power rating >500 kW) 4.7–125.8 Kubiszewski et al. [20]; n >40 8.9 8.1–34.5 (12.6) Dupont et al. [94] Kis et al. [74] (min-max)(base) Wind offshore 8.7 5.4–66.7 Carbajales-Dale [34]; n =37 14.8–51.3 Kubiszewski et al., [20]; n >4 12 6.9–19.1 (13.5) Dupont et al. [94] Kis et al. [74] (min-max)(base) Solar PV 7.8 8.7–34.2 Bhandari et al., [14]; n =23 7.2 2.7–7.5 (4.8) Dupont et al. [94] (present) Kis et al. [74] (min-max)(base) CSP 2.6 5.2–6.7 5.4–17.9 (9.8) Dupont et al. [94] (present) Kis et al. [74] (min-max)(base) 9.6–67.6 de Castro and Capellán-Pérez [25] n =13 5. Discussion We structure the Discussion section in two parts. First, a discussion on the potential role of future technological improvements to improve the EROI of the studied technologies in this study in the context of their likely large deployment in the next decades. Second, we give our views on how to properly take into account the effect of EROI for the energy transition and the potential implications we foresee. Energies 2020,13, 3036 18 of 42 5.1. On the Role of Future Technological Change The EROIs computed in this study correspond to estimates for current RES technologies and power plants in operation. Future technological improvements may reduce the EnUs and/or improve the CF of the different RES technologies, as it has been happening in the last decades (e.g., increased cell efficiency and reduced wafer thickness in solar PV [ 31 , 110 ]). Since RES are expected to increase their share substantially in the next few decades, it is very relevant to assess to what extent future technological improvements and increase in mineral recycling rates may contribute to increase their current EROI levels. In fact, the recycling of materials generally requires less energy than their extraction from the mines (virgin), which holds true as long as the recycling rates do not approach 100% due to thermodynamic limits. Particularly, in the future there will be some very relevant factors, which will tend to offset potential technical improvements as the renewables progressively scale-up at large levels and gain a substantial share in the energy mix: • Additional energy investments and losses related with variability management of RES such as storage capacity (PHS, electric batteries, hydrogen, etc.), power-to-X, curtailment, and additional grids. The first three factors will tend to lower the EROI of the system due to the Energy Stored On energy Invested (ESOI) of the storage device, the increase of the transformation phases with associated transformation losses, and/or the diminishing effective CF of power plant being electricity curtailed. The fourth factor includes the substantial adaptation and expansion of the existing grids to cope with connecting the new RES power plants as well as helping to evacuate power when unevenly produced and demanded. • The effect of decreasing returns in the potential of renewables, i.e., after best places are occupied it is necessary to move to more uneconomical sites [ 94 , 103 , 111 ], phenomena that may be worsened in some cases by land availability constraints [59,112]. • The increase in energy requirements for mineral processing related with ore grade decrease of minerals due to increased cumulated extraction [113–115]. •Thermodynamic limits to the continuous reduction of required energy investments (e.g., related with limits to substitution). • Limits to recycling rates (other than thermodynamic): as aforementioned, most of the machining processes require some virgin or pure material because the recycled scrap cannot be fully reused [90]. • The scarcity of some minerals in the future may drive the shift to more abundant minerals, which in turn are generally characterized by a lower performance (e.g., Ag instead of Al in mirrors, Nd in permanent magnets, Te in thin films, etc.) [63,86]. • Thermodynamic limits from the side of generation. For example, the fact that there are absolute limits to the height of rotors for wind or the Benz law (modern large wind turbines already achieve peak performance coefficients in the range of 45–50%, which is pretty close to the limit of 59.26% [ 94 ])), or the limits in the conversion from sunlight to electricity, such as the Schokley–Queisser limit for single-junction solar cells. Although the latter limit could be overcome with multi-junction technologies, the key general question is how realistic it is, considering that the most sophisticated technologies—also related with the previous point—are really scalable at a significant level compared with total energy demand, or if, in the future, they will rather remain marginal. Studies extrapolating the past evolution of EROI levels generally ignore all of the above factors and hence should be taken with care (e.g., [ 116 ]). Even studies that focus on past data are not totally reliable. For example, Louwen et al., [ 117 ], although they claim that all the PV systems analyzed include modules +inverter +mounting structure, in reality, the two more recent works (which seem to be largely driving their learning curve estimation, see their Figure 2b) correspond to (1) only PV modules, and (2) a solar rooftop system. Moreover, both of them use data from EU, not from China, where most production is located now. Energies 2020,13, 3036 19 of 42 Moreover, there is an additional issue related with the temporality of the transition, which will be especially relevant in the next few decades. Due to the fact that RES power plants, differently to FF ones, require large up-front costs and obtain delayed returns over the lifetime, a fast transition to RES would imply draining large shares of the energy available in order to sustain the energy transition. This means that, as shown in Capell á n-P é rez et al. [ 3 ], which in fact is based on many of the assumptions and data reported here the EROI of the full system could temporarily be well below the weighted average of the static EROI of the technologies and their supporting systems (e.g., grids, storage, etc.). Hence, at system level, the fast penetration of renewables can lead to a situation of “energy trap”, i.e., a reduction in the discretionary energy arriving to society simultaneously with the increase in the consumption of primary energy [ 3 , 118 ]. Under this case, the efficiency of the full system measured as the primary to final energy ratio would worsen. Since some definitions, inputs and parameters are different in this work with relation to those used in Capell á n-P é rez et al., [ 3 ] (here we perform a more detailed lifecycle analysis) in Appendix Ewe perform a harmonization of the results of both studies. The consistent comparison of the results obtained in both studies shows that the updated EROI st results presented in this work are lower (20–30%) for all the technologies, which reinforces the implications highlighted in the previous work [ 3 ] with relation to the potential scenario of “energy trap” during a fast energy transition. This phenomenon of “energy trap” would be aggravated in a context where population and energy consumption per capita are expected to continue increasing such as in the Green Growth paradigm. Hence, the relevant feature is not the future EROI over the lifetime of each specific technology, but that of the full energy system in a dynamic, transitional way. Hence, the ultimate objective is to assess the potential implications that the dynamic EROI over time of the full energy system might imply for the thriving future societies (i.e., its implications for income, employment, etc., but also for dimensions more difficult to capture in quantitative modeling such as diversity in social functions). In this sense, it should be taken into account that a “sufficiently high” EROI is a necessary but not sufficient condition to achieve sustainable energy systems while maintaining high complexity in society. In particular, it does not give information about key aspects of the different energy technologies such as their environmental impacts (CO 2 emissions, land occupation, etc.), social acceptance, future mineral availability, etc., as well as others such as cost effectiveness (although not totally being disconnected from energy effectiveness, cf. [ 119 ]). Although composite EROI indicators have been proposed to account for this (e.g., [ 60 ]), we believe that since sustainability is an inherent multi-dimensional concept its assessment has to be performed through a multi-dimensional set of indicators, being one of them the EROI. It is noteworthy that most models used for advising policy (e.g., IEA, IPCC, national governments, etc.) neglect the energy investments related with the construction and operation of the RES power plants, as well as the implications on the full energy system [ 55 , 118 , 120 ]. Among the few models considering this factor are models published in the scientific literature with, unfortunately, little (if any) political incidence, such as GEMBA [ 121 ]; NETSET [ 122 ]; EETRAP [ 118 ]). The relation of EROI to net energy is non-linear (i.e., the “net energy cliff”), and consequently its impact can potentially be misjudged. Given the metabolic implications of the variation of the EROI of the system, assuring the consistency between physical investments (energy and materials) and economic investments seems a key precondition to ensure the robustness and viability of alternative sustainability scenarios, such as the Green Growth [123–128] or Post-Growth [129] scenarios. We would like to conclude this section with a comment on the meaning of technological improvements and the relationship between economic (monetary) and biophysical (energetic and material) costs. Monetary cost reductions are typically identified with technological advances. For example , since the end of 2009, wind turbines prices have fallen by 30–40%, and solar PV module by around 80% [ 130 ] (and nearly 100 × between the 1950s and the mid-2000s, more than any other energy technology in that period [ 131 ]). The levelized cost of electricity of solar PV (utility scale) has been estimated to have fallen by more than 60% between 2010 and 2016 [ 130 ]. However, in reality Energies 2020,13, 3036 20 of 42 these historic cost reductions cannot be solely attributed to a reduction in the material and energetic intensities of a technology (technological improvement), given that they have also been critically affected by one-time financial (e.g., low interest rates to finance RES capital-intensive investments) and economic factors (e.g., economies of scale when increasing production, outsourcing to countries with less strict labor, environmental legislation, etc.). Additionally, the price of raw materials is subject to multiple influences (institutional framework, oligopolistic market structure, etc.) [ 132 – 134 ], which makes erratic their long-term evolution. For example, for the case of solar PV, it has been found that the reduction in average production cost and price of solar panels has been driven by factors, such as the reduction in the price of polysilicon, the increasing market penetration of lower cost firms from China, the increasing size of the facilities, and increases in industry investment, besides technological improvement, mainly in the form of the reduction in the use of polysilicon, and improvement of panel efficiencies [110,131]. Hence, a direct relationship between monetary costs and energy costs, or technological improvement, does not exist, although of course they are not totally independent [ 27 , 119 ]. Despite the aforementioned empirical evidences, the most common method used to represent endogenous technical change in energy–economy and integrated assessment models that inform energy planning and policy analysis is based on the extrapolation of historical learning curves (also known as experience curves) [ 2 , 130 , 131 , 135 ], which are a log–linear equation derived from empirical observations relating the unit cost of a technology to its cumulative installed capacity or electricity generated. A simpler version known as Moore’s Law treats time as the independent variable. This model is sometimes complemented by relating unit cost to cumulative expenditures for research and development, but it is, in any case, built under the assumption of perpetual cost reductions linked to increased production assuming that growth increases the likelihood of fundamental technological advances, incremental learning by doing, economies of scale in manufacturing, and standardization. However, in the era of a “full world” and over-exploited and degraded biosphere (“the global economy is now so large that society can no longer safely pretend it operates within a limitless ecosystem“ [ 136 ]), these dynamics cannot anymore be taken as granted in the future, which represents a paradigmatic change with relation to the past decades. “Natural resource flows are now the scarce factor, and labor and capital stocks are now relatively abundant. This basic pattern of scarcity has been reversed by a century of growth” [136]. 5.2. Implications of Taking into Account the EROI for the Transition to RES The first intuitive answer about the implications of taking into account the EROI of the RES during the energy transition can be given by the “net energy cliff”. Figure 6shows the current global average EROI of each RES technology for electricity generation at the three boundaries studied in this study, represented as a share of the resulting share of net power vs. gross power in the net energy cliff. It can be seen that below levels of EROI <~5–3:1, the share of net energy declines abruptly. As aforementioned in the introduction, one of the open debates of greatest importance today in the context of the energy transition is whether the RES with higher potential have a sufficient EROI to maintain the energy “metabolism” of a large and complex civilization such as the industrial one. Given that, the estimation of a minimum EROI to maintain complex societies is a complex, and to date, elusive task involving both technical and societal issues, a large part of the literature has focused on the comparison with the EROI of FF as an indirect way of shedding light on this. In fact, in the case that RES were to be found to have an EROI level, at least as high as the one characterizing FF, then it could be assured that RES could also fuel complex modern societies. Energies 2020,13, 3036 21 of 42 Energies 2020, 13, x FOR PEER REVIEW 22 of 43 0% 20% 40% 60% 80% 100% 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 Discretionary energy (% of final) EROIext RES for electricity generation in the net energy cliff (current global-average EROI ext ) Large hydro Wind onshore Wind offshore PV CSP Batteries 0% 20% 40% 60% 80% 100% 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 Final energy (% of primary) EROIst RES for electricity generation in the net energy cliff (current global-average EROI st ) Large hydro Wind onshore Wind offshore PV CSP Batteries 0% 20% 40% 60% 80% 100% 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 Net energy (% of final) EROIfin RES for electricity generation in the net energy cliff (current global-average EROI fin ) Large hydro Wind onshore Wind offshore PV CSP c a b EROIFFext_elec Brockway et al., (2019) Figure 6. Current global average EROI of each RES technology for electricity generation at ( a ) standard; ( b ) final; ( c ) extended level represented as a share of the resulting share of net power vs gross power in the net energy cliff. CSP is not represented in figure ( c ) since its EROI ext <1:1. The figure ( c ) also shows the approximated value of 3.9:1 (black dotted line) for the global-average EROI ext for fossil fuels-fired electricity re-estimated from Brockway et al. [ 16 ] to adapt of our methodology (see Appendix Dfor the details). Energies 2020,13, 3036 22 of 42 However, the comparison of the EROI of RES at the extended boundary with the one of FF is far from being straightforward. To the knowledge of the authors, Brockway et al., [ 16 ] is the only study having computed the global average EROI ext of FF at the electricity generation level. However, their input–output extended method is unsuitable for computing the EROI of RES given that it does not account for the energy associated to the capital investments (neither new capacities, decommission, nor new transmission lines required to move from EROI standard to point-of-use level). Although capital investments may represent a reduced part of the energy requirements for the case of an already extensively deployed industry, where most of the energy inputs are required in the O&M phase (as it can be assumed, not being so far from the situation of the current FF industry globally), for RES, the situation is exactly the opposite. Most energy investments are made upfront and the installed capacity is growing globally very fast. Although perfect comparability between both studies is not possible, we have made an effort to approximate the EROI ext (gross) computed by Brockway et al., [ 16 ] for the aggregated FF to be roughly comparable with our method. When roughly accounting for the most relevant factors not considered in their study, we obtain a value of 3.9:1 (see Appendix D), which allows to conclude that their reported result of ~4:1 (gross EROI ext ) seems robust. Hence, it can be concluded that, very likely, the global average EROI ext of variable RES is currently lower than those of FF for electricity generation. This means that the current substantial lower power density (MWh/m 2 ) [ 59 , 137 ] and higher material intensity (kg/MWh) [ 4 , 5 ] of variable RES, with relation to FF, make that the initial energy investment of RES weights more than the energy savings, due to the very low O&M energy requirements phase with relation to FF, where the situation is rather the opposite. The case of hydropower can be regarded as an exception given its very long lifetime (3–4 times for variable RES) and that although its power density is rather low, most parts of the surface is occupied by water “naturally” contained by valleys and mountains, and the built infrastructure is just concentrated in the dam. Hence, of those RES with a higher techno-sustainable potential, the Figure 6c shows that only large hydroelectricity has currently a high EROI ext clearly above the current global-average FF for electricity generation. Given that the techno-sustainable potential of large hydro is limited to less than double the current installed capacity (e.g., [ 138 ]), the transition to RES to supply the expected (increasing) energy demands will require large shares of the variable RES; hence, driving the EROI of the system to lower values, especially, as aforementioned, during the transition period before reaching a more stationary situation. Hence, to avoid driving the energy system to excessively low EROI levels, it is key that those uses that cannot be supplied by electricity, such as heat, are instead supplied directly by thermal technologies such as solar thermal, geothermal, and bioenergy in order to minimize energy conversions. Hence, there is a trade-offbetween the low EROI of RES for electricity with the strategy of managing variability through increasing the interconnectedness of the heat, cooling, transport, and electricity sectors [ 139 ] given that the latter introduces more conversions (i.e., losses) in the system. However, as previously highlighted, specific regional conditions are a key factor to take into account when designing the transition towards sustainable energy systems fully based on RES. The above comparison holds for the current situation. However, it has to be taken into account that, in the future, there will be three concurrent dynamics that will affect the EROI of the system with different sign: • The EROI from FF globally is in a decreasing trend [ 16 , 19 , 44 , 45 ], a trend which will be intensified due to the higher share of unconventional fuels [19,140,141]. • Technological improvement of RES (and in general any factor going in the direction of EROI max in the uncertainty analysis reported in Appendix C). • The likely large scale deployment of RES due to the enforcement of sustainability policies will (1) tend to replace FFs in the energy mix, and (2) drive an array of factors overviewed in the previous Section 5.1, which nowadays are negligible, but will become increasingly important as the renewables progressively scale-up at large levels and gain a substantial share in the energy mix. These factors will put a downward pressure on future EROI. Energies 2020,13, 3036 23 of 42 Future work will be directed to replicate the present study to compute the EROI of FF at the three EROI boundaries in order to be able to perform a consistent comparison. In any case, if globally the EROI from FF continues to fall, as it seems to have been doing in the recent past, there will be a point where FF will not be able to maintain complex industrial societies in the future, even if there are still resources in the Earth’s crust due to delivering insufficient energy surplus. This dynamic will be in reality aggravated by the fact that the environmental impacts produced by the burning of these FF (e.g., climate change impacts) will also contribute to hamper greatly our societies [1]. 6. Conclusions In this work, we have estimated the current global-average EROI st (standard farm-gate), EROI final (consumer point-of-use), and EROI ext (extended, including indirect investments), for the five RES energy generation technologies assessed to have the highest techno-sustainable potential for electricity generation (solar PV and CSP, wind onshore, wind offshore, and large hydroelectricity). The same methodology, based on an extensive and comprehensive literature review, is applied in order to collate data about the material requirements for the power plants’ construction and operation of each technology, which allows for an internally consistent inter-comparison between the EROI levels of different technologies. The obtained current global average EROI for each technology of electricity generation at the three boundary levels (standard, final, and extended, respectively) are large hydro (28.4, 13, 6.5:1) > wind onshore (13.2, 5.8, 2.9:1) >wind offshore (8.7, 4.7, 2.3:1) >solar PV (7.7, 3.5, 1.7:1) >CSP (2.6, 1.6, 0.8:1). Hence, accounting for the transmission losses, the energy investments associated to the grids as well as the change from plant to full-system context implies a large gap between the three EROI boundary levels studied, a gap that other authors have also found for FFs (e.g., [16,33]). The extended boundary encompassing the implications for the full-system is the relevant perspective to assess the contribution in terms of energy surplus for discretionary uses to the society (e.g., growing food, family support, education, health, culture, etc.) of an energy technology. An array of factors make that, to maintain a sustainable modern complex society, the EROI ext of the energy generation technologies should be distinctly greater than 1:1: a sufficient diversity of jobs and enterprises [ 38 , 39 , 58 ], the need to minimize the environmental impacts [ 60 ], the need of a “security buffer” to overcome unexpected events, such as accidents or natural disasters, and last but not least, the fact that most current societies are very unequal and, thus, are diverting energy towards metabolically “useless” purposes, such as those caused by inequalities. However, few works have, to date, dealt with the intricate issue of the minimum EROI to sustain modern complex societies, and none has been conclusive to date at extended level. When comparing with FF, our results indicate that the global average EROI ext of variable RES is currently lower than the one corresponding to the global-average FF for electricity generation (4:1, reported by Brockway et al. [ 16 ], which adjusted to our methodology for comparability, corresponds to roughly 3.9:1). Large hydroelectricity stands out with a high EROI ext of ~6.5:1 due to its particular features, while CSP stands in the opposite side with an EROI ext <1:1 meaning that this technology currently works rather as a storage device than an energy generation one. However, the heterogeneity of the EROI of variable RES at the geographical level is also a relevant feature to take into account, as it has been showed for the case of solar PV. Our analysis is a static analysis. However, the implications of the EROI of RES must be assessed dynamically. During the forthcoming transition to RES, different dynamics will be simultaneously in action: declining EROI trends from FF, technological improvement of RES, and progressive emergence of an array of factors that, nowadays, are negligible, but which will become increasingly important as the renewables progressively scale-up at large levels and gain a substantial share in the energy mix: RES variability management, decreasing returns in the potential of renewables [ 94 , 103 , 111 ], the increase in energy requirements for mineral processing related with ore grade decrease of minerals due to increased cumulated extraction [ 113 – 115 ], thermodynamic limits to the continuous reduction of required energy investments (e.g., related with limits to substitution), limits to recycling rates, Energies 2020,13, 3036 24 of 42 the scarcity of some minerals in the future [ 63 , 86 ], and thermodynamic limits from the side of generation. Additionally, the temporary fast growth of RES sources and the dismantling of the ones they replace (FF) will temporally reduce the EROI of the full system well below the weighted average of the static EROI of the technologies and their supporting systems (e.g., grids, storage, etc.). A shown in previous work [ 3 ], a fast energy transition with a very rapid growth of variable RES can generate issues at system level (energy trap). The present work, including a more detailed life-chain analysis, results in lower current EROI st levels than the ones used in [ 3 ] as inputs for the dynamic energy transition analysis (see Appendix E), hence, reinforcing the conclusions presented there. Noteworthy, the latter phenomena will be aggravated in a society where population and energy per capita have to continue to grow. Hence, the design of viable sustainable energy systems must take into account the dynamic evolution of the EROI of the system, especially for modern societies pursuing continuous economic growth which otherwise may derive in unintended “energy-trap” scenarios. This work documents in detail the assumptions and data used to estimate the energy and material investments to install, operate, and maintain the RES with higher potential in the MEDEAS modeling framework [ 3 , 55 , 56 ]. In this framework, the computation of the “static” (over the lifetime) EROI of the different energy technologies is an intermediate step to consistently integrate the mineral and energy requirements of the energy transition in a wider modeling framework able to represent the dynamic nature of the transition. Further work will focus on the estimation of the mineral and energy requirements for the rest of energy technologies, which will allow the comprehensive computation of the variation of the EROI of the full energy system in scenarios of energy transition. A step further will assure the consistency with the monetary investments, which will allow to ultimately capture the implications for macroeconomic and socioeconomic dimensions, such as the change in the economic structure or the available income to households. This approach may ultimately allow to approach the issue of the minimum EROI to sustain complex modern societies from an endogenous point of view, taking into account both technical options and societal decisions, which could restrain or boost the vicious circle dynamics of declining EROI. Supplementary Materials: The following are available online at http://www.mdpi.com/1996-1073/13/12/3036/s1. Author Contributions: Conceptualization, C.d.C.; methodology, C.d.C.; validation, I.C.-P.; formal analysis, C.d.C.; investigation, C.d.C. and I.C.-P.; resources, C.d.C. and I.C.-P.; data curation, C.d.C. and I.C.-P.; writing—original draft preparation, C.d.C.; writing—review and editing, I.C.-P.; visualization, I.C.-P. All authors have read and agreed to the published version of the manuscript. Funding: This work has been partially developed under the MEDEAS and LOCOMOTION projects, funded by the European Union’s Horizon 2020 research and innovation programme under grant agreements no 691287 and 821105, respectively. The authors are thankful as well for the support of MODESLOW (Modeling and Simulation of scenarios towards a LOW-carbon transition: The Spanish case), a Spanish national research project funded under the Spanish National Research, Development and Innovation Program (Ministry of Economy and Competitiveness of Spain, ref. ECO2017-85110-R). Iñigo Capell á n-P é rez also acknowledges financial support from a Juan de la Cierva Research Fellowship of the Ministry of Economy and Competitiveness of Spain (no. FJCI-2016–28833). Acknowledgments: Authors thank the Group of Energy, Economy, and Dynamics Systems (GEEDS) of the University of Valladolid for indirectly contributing to this work during group discussions. Conflicts of Interest: The authors declare no conflict of interest. Appendix A. EnUs The Supplementary Material excel spreadsheet collates the materials required for the construction and operation of the power plants of the different technologies per MW (kg/MW). For each material, the embodied energy is given in primary terms (MJ/kg) using current recycling rates (Equation (7) of the main text). For the calculation of the denominator of the EROI (Equations (4)–(6)), we distinguish four EnU components: Total EnUdirect: EnUNew cap +EnUO&M +EnUDecom wear cap +EnUG&S (A1) Energies 2020,13, 3036 25 of 42 where: EnU G&S is calculated after the information in Appendix B(see also the tab “Electric grids” in the Supplementary Material). EnUDecom =0.1·EnUNew cap EnU O&M is estimated annually; for solar technologies the material requirements and energy intensities are in the “Solar PV & CSP” tab, for wind technologies the EnU O&M is assumed to be 1% and 1.25% for onshore and offshore relative to their EnU New cap , respectively, as explained in the “Wind onshore & offshore” tab of the Supplementary Material. Finally, EnUNew cap is divided in three components: EnUNew cap =EnUNew cap CtoG +EnUNew cap GtoE +EnUNew cap GtoG (A2) Cradle to first gate (CtoG) is the part of a complete cycle analysis between the mineral mines extraction to the basic industrial material like rods, plates, etc., used by the industry. Gate to gate (GtG) represents the manufacturing phase: the process between rods, plates, etc. and the final form of these materials in a power plant (e.g., computer, frame of a PV plate, rotor, etc.). EnUNew cap CtoG is the embodied energy in the Raw/resource to the Suppliers chain (cradle to gate in LCA terminology), EnUNew cap GtoG is the embodied energy from the suppliers to the finished products of the power plant or the Manufacturing phase (gate to gate phase), and EnUNew cap GtoE is the embodied energy in the Erection of the power plant. Both, EnUNew cap CtoG and EnUNew cap GtoE are calculated applying the Equation (7) to the data collated in the Excel spreadsheet for each item. As explained in the main text, there is not enough detailed data of the Manufacturing phase EnUNew cap GtoG for RES technologies, and we have assumed that this phase represents a fixed share for each technology of the Suppliers and Erection phases: EnUNew cap GtoG =x·EnUNew cap CtoG +EnUNew cap GtoE (A3) As stated in the main text, taking into consideration the machining energy and the % of scrap in the manufacturing phase, x has been approximated to 0.25 for solar CSP (more precisely the calculation would be [1−(1−0.15) ×(1−0.10)] =0.235) for wind onshore and wind offshore, to 0.20 for solar PV. Cumulative Energy Demand (CED) is a term with origin in the life-cycle analysis (LCA) community, where it is defined including all the primary energy harvested in the lifetime. However, this definition needs corrections to calculate EROI or the Energy Payback time. To avoid confusion of the different “CEDs” being used in the literature, and given priority to the historical precedence to the CED defined by LCA community, we apply in this paper the term Energy Used (EnU) instead of Cumulative Energy Demand (CED) (the same criteria was applied in precedent works of the authors [3,25]). Appendix B. Material Requirements and Performance Factors Per Technology Details Solar CSP (CSP with Molten Salt Storage Without Backup) Performance Factors: De Castro and Capell á n-P é rez [ 25 ] showed that the current CF of CSP globally is <0.25; we use CF =0.252 based in IRENA [ 93 ] for the sake of using the same reference for the different studied technologies. We take 25 years of lifetime [ 25 ]. SC =9.1% is the 2014–2018 annual average from [ 95 – 99 ] (Kis et al. [ 74 ] give 15% of “parasitic loads” for CSP plants with storage. Kis et al., [ 74 ] give 15% of “parasite loads” for CSP plants with salt storage and 7.2% for no storage plants. Ramorakane and Dinter [ 142 ] give approximately also 15% for the power plant Andasol 3. The parasitic load (PL) is the self-consumption from the power generated but not dispatched plus the imported power from the grid. This imported power is at least 50% because during night the PL is around 50% of the maximum during the day. Ramorakane and Dinter do not give results for other auxiliary facilities like domestic Energies 2020,13, 3036 32 of 42 Appendix D.1. Factors Tending to Overestimate the Reported EROIext of FF in Brockway et al. (2019) We identified three relevant factors tending to overestimate the reported EROI ext of FF in Brockway et al. [16]: 1. No computation of energy investments related with the construction of energy facilities. As Brockway et al. [ 16 ] themselves point out (page 620): “our EROI estimates do not include any energy invested in the production of energy associated with the fixed capital equipment in the energy production [and transportation] industries . . . [our EnU estimates] are very likely to be underestimated and should be considered as lower-bound values . . . ”. 2. No computation of the decommissioning phase. 3. Not fully capturing the energy investments associated to the grids. The two first factors prevent in fact the estimation in that work [ 16 ] of the EROI of RES, even though they conclude similar or better values of EROI final of RES technologies than FF through indirect comparison with other studies. Additionally, the third factor, as argued in the main text, could be more relevant for RES than FF. Appendix D.2. Factors Tending to Underestimate the Reported EROIext of FF in Brockway et al. (2019) Three other relevant factors are identified tending to underestimate the reported EROI ext of FF in Brockway et al., [16]: 1. No computation of heat in CHP plants: when the EROI of FF power plants is computed, it should also be taken into account that FF power plants can also generate commercial heat. 2. Not taking into account that the “own use” (embodied energy in the denominator of the EROI) dedicated to final “non-energy uses” of the FF primary sources (plastics, lubricants, etc.) is not attributable to the energy system because it is not energy for energy but energy for matter. Plastics, lubricants, etc., are used by the entire industry and the rest of the economy and even for the RES industry, their embodied energy use must be attributed case by case. The quantity of “non-energy uses” that reenters the FF economy is minute relative to the entire economy, and must be discounted not added. Moreover, some of this matter (e.g., plastics) could even reenter the energy system at the end of their lifetime (e.g., electricity from “waste”). 3. Not all of the FF own-energy use can be attributed to producing FF energy: following Table 2 in [ 16 ], the direct EnU includes all the FF own-energy use directly used by the respective industries of coal, gas, and oil; hence, in the extraction, refining and conversion to FF-derived final fuels. However, a significant part of FF are used to produce energy that are used by other energy technologies, such as the diesel used for the construction of a RES power plant that must be attributed to the RES power plant and also their embodied energy in the refinery and the oil extraction, etc. Hence, the parameters shown in Tables 2 and 3 of [16] seem overestimated. In order to be able to compare the FF-global-average fired-electricity EROI ext with the EROI ext level obtained for RES in this work, in this appendix we adjust the obtained results by Brockway et al., [ 16 ] taking into account the under/over estimation factors described above and translate them to our methodology. Table A2 reports the parameters and performance factors assumed for FF power plants, which allow to re-estimate the EROI ext of global-average FF for fired-electricity applying our methodology. Energies 2020,13, 3036 33 of 42 Table A2. Parameters and performance factors assumed for FF power plants in order to compare Brockway et al. [16] with our results for RES power plants. Parameter/Performance Factor Value 1 Capacity factor (CF) 0.45 2 Lifetime (L) 45 years 3 Operational Losses (OL) 5% 4 Transmission and Distribution Losses (TDL) 9.2% 5 Final electricity output (1 MW power plant capacity) 55.09 ×107MJ/MW 6 EnU in Brockway et al. (O&M direct +indirect) 13.77 ×107MJ/MW 7 EnU in O&M of grids (direct +indirect) 4.14 ×107MJ/MW 8 EnU in construction +decommissioning phases (direct +indirect) 1.38 ×107MJ/MW 9 Commercial heat output (corrected for quality) 6.25 ×107MJ/MW 10 Own use to non-energy uses (plastics, etc.) 1.38 ×107MJ/MW 11 Own use dedicated to non-FF energy sources 2.18 ×107MJ/MW 1. CF: own estimation based in the IEA Sankey [72] and total power plants capacity [159]. 2. L: own estimation based in [160,161]. 3. OL: rough own estimation based in [74] from the average losses of fossil fuel plants. 4. TDL: own results, see main text. 5. Final electricity output: assuming the former parameters and applying eq. 5 of the main text {= 31.54 ×106×0.45 ×45 ×(1−0.05) ×(1−0.092)} for 1 MW power plant during their lifetime. 6. EnU in Brockway et al. (O&M direct and indirect): Following Brockway et al. [ 16 ] results, the EROI ext of FF power plants is 4:1; therefore, assuming the final electricity output (5th point here) this translates to 13.77 ×107MJ/MW. 7. EnU in O&M of grids: in the main text we have estimated 1.15 × 10 7 MJ/MW for the direct O&M of grids during 25 years, as the lifetime of the FF power plants is 45 years this amounts to 45/25 times more direct costs. Following our methodology, we have assigned the same embodied costs for the indirect costs; therefore, the total EnU in O&M of grids will be: 2 × 1.15 × 10 7× 45/25. 8. EnU in C +D phases: we estimate for the direct costs a 5% of the total EnU accounted in Brockway et al., [ 16 ]. Kis et al. [ 74 ] results for FF power plants are less than 5% for this two phases over the total in their direct embodied energy (LCA based). Assuming another 5% for indirect costs, we arrive to a 10% of the denominator in the O&M phase accounting of Brockway et al. methodology. Then we take 10% of this number to account for this C +D phases. 9. Commercial heat output: 15% is the commercial heat from power plants relative to their electricity output at plant phases (estimation based in the IEA Sankey in 2015 [ 72 ]). This output could be corrected by a quality factor, following the criteria of “final to primary” factor “g” as the “system quality of the energy mix”. We multiply this heat output by the factor g (=0.688) that we have used in our methodology (see main text). This result to 6.25 × 10 7 MJ/MW of “corrected” final energy output due to commercial heat (at plant phase the electricity output is 6.06 × 10 8 MJ/MW, then 15% of that number multiplied by “g” is the “heat” output to be added to the electricity output ). 10. Own use to non-energy uses: the “own use” of the power plants that is inverted to “non-energy uses” and not to the energy sector (embodied energy to fabricate materials, such as plastics, lubricants, asphalt, fertilizers, etc., which are products of the FF industry, must not be assigned to the denominator of the FF power plants, other than the plastics, lubricants, etc., that are used in this industry, which will be minute in comparison with the entire economy). We estimate that at least 9.1% of the “own use” of the denominator of the EROI ext is not an embodied energy attributable to the FF because this is the proportion of the non-energy uses relative to the final energy (IEA Sankey 2017 data [ 72 ]). In fact, the fabrication of most “non-energy uses” are much more energy intensive that the oil to diesel in the refinery process (e.g., plastics, fertilizers). Moreover, some of the material outputs of FF economy at the end of their lifetime could reenter Energies 2020,13, 3036 34 of 42 the energy economy in the form of “waste power plants”. Furthermore, the economy of FF is using, indirectly, the high-embodied energy in the chemical bonds of FF that is incorporated in the material products; a hypothetical substitution of these products (plastics, lubricants, fertilizers, etc.) will likely need more energy to fabricate alternatives, and less energy for discretionary uses of energy will have. This difficults the comparison between FF and RES power plants at this extended and system level vision. We take 10% of the denominator of Brockway et al. [ 16 ] (10% of 1.38 ×108MJ) as a lower bound guess. 11. Own use of FF dedicated to other non FF energy sources: the relative proportion of the electricity own use (Tables 2 and 3 from Brockway et al. [ 16 ]) that is used not for the FF power plants but for the EnU of the rest of the power plants (nuclear, RES). This amount must be attributable to the rest of energy sources and not to fossil fuels (in the EnU of RES of our main text are attributed to RES and not to FF). Non-FF sources of energy account for 15.8% of the primary energy (non-energy uses discounted) following the Sankey 2017 of IEA [ 72 ] (in final terms assuming only a 33% efficiency for biopower and nuclear the result is 20.1% for the contribution of non FF). This 15.8% must be attributed to these non-FF sources. Then the denominator of Brockway et al. [ 16 ] is overestimated by around 0.158 ×1.38 ×108MJ/MW. From Table A2, we can re-estimate the EROI ext of global-average FF for fired-electricity applying our methodology as follows: EROIext FF (adjusted to our methodology) =(55.09 +6.25)/(13.77 +4.14 +1.38 −1.38 −2.18) =3.9:1. (A5) It is noteworthy the importance of grids also for FF, accounting for 30% of the O&M costs estimated by Brockway et al. [ 16 ]. Therefore, very roughly, the estimation of Brockway et al. [ 16 ] holds because the factors tending to overestimate and underestimate the EROI following our methodology seem to tend to cancel out. Appendix E. Comparison of Results with Capellán-Pérez et al. (2019) Capell á n-P é rez et al. [ 3 ] computes the dynamic EROI of the global energy system associated to different penetration levels of RES for electricity by 2060 accounting for the material and energetic inputs of the variable RES (CSP, PV, wind onshore, and wind offshore), dispatchable RES and storage systems (electric batteries and PHS). They conclude that the EROI st of the system could fall below the threshold of 5–10:1 which has been speculated in the literature to be able to support complex modern societies, at least during a fast transition to renewables as the one we should perform in the next few decades to avoid dangerous climate change. In the present work, we have performed, for these four technologies, a more in-depth analysis of the material and energy requirements and parameters and we have extended the boundaries from the EROIst to the EROIfinal and EROIext. However, the comparison between the present study and the previous one in [ 3 ] is not direct since the definitions used in each of the work for the estimation of the EROI st are different (see Section 3. Methodology). This Appendix Eis, hence, directed to bridge the gap between both studies and report the EROIst values computed in [3] translated to the definition and with the performance factors used in the present work. Table A3 compares the “static” EROI st levels obtained using the input data, EROI definition and performance factors from Capell á n-P é rez et al. [ 3 ] (results not reported there) (“from [ 3 ]”), with the results using the same input data used in [ 3 ], but with the EROI equation and performance factors used here (“Adaptation of [ 3 ]’s results . . . ”), and with the results obtained in this work (“Present work”). Energies 2020,13, 3036 35 of 42 Table A3. Comparison of the EROIst levels estimated in the present work and from Capellán-Pérez et al. [ 3 ]. The first line reports the original results while the second is adapted to the EROI definition and performance factors used in this work and is then comparable with the third line. EROIst Wind onshore Wind offshore Solar PV Solar CSP From [3] 16.0 9.8 8.2 3.7 Adaptation of [ 3 ]’s results to the EROI definition and performance factors used in this work 16.2 12.2 9.6 3.3 Present work (see Table 3) 13.2 8.7 7.8 2.6 Table A3 shows that the updated EROI st results are lower (20–30%) for all the technologies, which reinforces the implications highlighted in the previous work [ 3 ] with relation to the potential scenario of energy trap during a fast energy transition (see discussion and conclusions in the main text). Abbreviations BAU Business as usual scenarios CF Capacity factor CSP Concentrated Solar Power EnU Energy used by the energy system to deliver energy EROI Energy Return On energy Invested ESOI Energy Stored On energy Invested FF Fossil fuels g quality correction factor of energy GG Green Growth scenarios HVDCs High-voltage, direct current Hydro Hydroelectric power plants IEA International Energy Agency IO Input–Output LCA Life-cycle analysis O&M Operation and maintenance OL Operational Losses PV Photovoltaic RC Recycled content ratio RES renewable energy sources SC Electricity self-consumption of the facility TDL Transmission and distribution losses References 1. IPCC. Global Warming of 1.5 ◦ C. Intergovernmental Panel on Climate Change (IPCC). Available online: http://www.ipcc.ch/report/sr15/(accessed on 9 June 2020). 2. IPCC. Climate Change 2014: Mitigation of Climate Change; Cambridge University Press: Cambridge, UK, 2014. 3. Capell á n-P é rez, I.; de Castro, C.; Miguel Gonz á lez, L.J. Dynamic Energy Return on Energy Investment (EROI) and material requirements in scenarios of global transition to renewable energies. Energy Strategy Rev. 2019 , 26, 100399. [CrossRef] 4. de Koning, A.; Kleijn, R.; Huppes, G.; Sprecher, B.; van Engelen, G.; Tukker, A. Metal supply constraints for a low-carbon economy? Resour. Conserv. Recycl. 2018,129, 202–208. [CrossRef] 5. Tokimatsu, K.; Wachtmeister, H.; McLellan, B.; Davidsson, S.; Murakami, S.; Höök, M.; Yasuoka, R.; Nishio, M. Energy modeling approach to the global energy-mineral nexus: A first look at metal requirements and the 2◦C target. Appl. Energy 2017,207, 494–509. [CrossRef] 6. Valero, A.; Valero, A.; Calvo, G.; Ortego, A. Material bottlenecks in the future development of green technologies. Renew. Sustain. Energy Rev. 2018,93, 178–200. [CrossRef] Energies 2020,13, 3036 36 of 42 7. Campbell, C.J.; Laherrère, J. The end of cheap oil. Sci. Am. 1998,278, 60–65. [CrossRef] 8. Capell á n-P é rez, I.; Mediavilla, M.; de Castro, C.; Carpintero, Ó .; Miguel, L.J. Fossil fuel depletion and socio-economic scenarios: An integrated approach. Energy 2014,77, 641–666. [CrossRef] 9. Creutzig, F.; Ravindranath, N.H.; Berndes, G.; Bolwig, S.; Bright, R.; Cherubini, F.; Chum, H.; Corbera, E.; Delucchi, M.; Faaij, A.; et al. Bioenergy and climate change mitigation: An assessment. GCB Bioenergy 2014 . [CrossRef] 10. Deng, Y.Y.; Haigh, M.; Pouwels, W.; Ramaekers, L.; Brandsma, R.; Schimschar, S.; Grözinger, J.; de Jager, D. Quantifying a realistic, worldwide wind and solar electricity supply. Glob. Environ. Chang. 2015 ,31, 239–252. [CrossRef] 11. IPCC. Special Report on Renewable Energy Sources and Climate Change Mitigation; Edenhofer, O., Pichs-Madruga, R., Sokona, Y., Seyboth, K., Matschoss, P., Kadner, S., Zwickel, T., Eickemeier, P., Hansen, G., Schlömer, S., et al., Eds.; Cambridge University Press: Cambridge, UK; New York, NY, USA, 2011. 12. Moriarty, P.; Honnery, D. What is the global potential for renewable energy? Renew. Sustain. Energy Rev. 2012,16, 244–252. [CrossRef] 13. Lambert, J.G.; Hall, C.A.S.; Balogh, S.; Gupta, A.; Arnold, M. Energy, EROI and quality of life. Energy Policy 2014,64, 153–167. [CrossRef] 14. Bhandari, K.P.; Collier, J.M.; Ellingson, R.J.; Apul, D.S. Energy payback time (EPBT) and energy return on energy invested (EROI) of solar photovoltaic systems: A systematic review and meta-analysis. Renew. Sustain. Energy Rev. 2015,47, 133–141. [CrossRef] 15. Brand-Correa, L.I.; Brockway, P.E.; Copeland, C.L.; Foxon, T.J.; Owen, A.; Taylor, P.G. Developing an Input-Output Based Method to Estimate a National-Level Energy Return on Investment (EROI). Energies 2017,10, 534. [CrossRef] 16. Brockway, P.E.; Owen, A.; Brand-Correa, L.I.; Hardt, L. Estimation of global final-stage energy-return-on- investment for fossil fuels with comparison to renewable energy sources. Nat. Energy 2019 ,4, 612. [CrossRef] 17. Dale, M. A Comparative Analysis of Energy Costs of Photovoltaic, Solar Thermal, and Wind Electricity Generation Technologies. Appl. Sci. 2013,3, 325–337. [CrossRef] 18. de Castro, C.; Carpintero, Ó .; Frechoso, F.; Mediavilla, M.; de Miguel, L.J. A top-down approach to assess physical and ecological limits of biofuels. Energy 2014,64, 506–512. [CrossRef] 19. Hall, C.A.S.; Lambert, J.G.; Balogh, S.B. EROI of different fuels and the implications for society. Energy Policy 2014,64, 141–152. [CrossRef] 20. Kubiszewski, I.; Cleveland, C.J.; Endres, P.K. Meta-analysis of net energy return for wind power systems. Renew. Energy 2010,35, 218–225. [CrossRef] 21. Prieto, P.A.; Hall, C.A.S. Spain’s Photovoltaic Revolution: The Energy Return on Investment, 2013th ed.; Springer: Berlin, Germany, 2013; ISBN 1-4419-9436-X. 22. Raugei, M. Comments on “Energy intensities, EROIs (energy returned on invested), and energy payback times of electricity generating power plants”—Making clear of quite some confusion. Energy 2013 ,59, 781–782. [CrossRef] 23. Weißbach, D.; Ruprecht, G.; Huke, A.; Czerski, K.; Gottlieb, S.; Hussein, A. Energy intensities, EROIs (energy returned on invested), and energy payback times of electricity generating power plants. Energy 2013 ,52, 210–221. [CrossRef] 24. Carbajales-Dale, M. When is EROI Not EROI? Biophys. Econ. Resour. Qual. 2019,4, 16. [CrossRef] 25. de Castro, C.; Capell á n-P é rez, I. Concentrated Solar Power: Actual Performance and Foreseeable Future in High Penetration Scenarios of Renewable Energies. Biophys. Econ. Resour. Qual. 2018,3, 14. [CrossRef] 26. Ferroni, F.; Hopkirk, R.J. Energy Return on Energy Invested (ERoEI) for photovoltaic solar systems in regions of moderate insolation. Energy Policy 2016,94, 336–344. [CrossRef] 27. Hall, C.A.S.; Klitgaard, K.A. Energy and the Wealth of Nations: Understanding the Biophysical Economy; Springer: New York, NY, USA, 2012; ISBN 978-1-4419-9398-4. 28. Koppelaar, R.H.E.M. Solar-PV energy payback and net energy: Meta-assessment of study quality, reproducibility, and results harmonization. Renew. Sustain. Energy Rev. 2017,72, 1241–1255. [CrossRef] 29. Murphy, D.J.; Carbajales-Dale, M.; Moeller, D. Comparing Apples to Apples: Why the Net Energy Analysis Community Needs to Adopt the Life-Cycle Analysis Framework. Energies 2016,9, 917. [CrossRef] 30. Palmer, G.; Floyd, J. An Exploration of Divergence in EPBT and EROI for Solar Photovoltaics. Biophys. Econ. Resour. Qual. 2017,2, 15. [CrossRef] Energies 2020,13, 3036 37 of 42 31. Raugei, M.; Sgouridis, S.; Murphy, D.; Fthenakis, V.; Frischknecht, R.; Breyer, C.; Bardi, U.; Barnhart, C.; Buckley, A.; Carbajales-Dale, M.; et al. Energy Return on Energy Invested (ERoEI) for photovoltaic solar systems in regions of moderate insolation: A comprehensive response. Energy Policy 2017 ,102, 377–384. [CrossRef] 32. Murphy, D.J.; Hall, C.A.S.; Dale, M.; Cleveland, C. Order from Chaos: A Preliminary Protocol for Determining the EROI of Fuels. Sustainability 2011,3, 1888–1907. [CrossRef] 33. Raugei, M. Net energy analysis must not compare apples and oranges. Nat. Energy 2019,4, 86. [CrossRef] 34. Carbajales-Dale, M. Chapter 21—Life Cycle Assessment: Meta-analysis of Cumulative Energy Demand for Wind Energy Technologies. In Wind Energy Engineering; Letcher, T.M., Ed.; Academic Press: Cambridge, MA, USA, 2017; pp. 439–473. ISBN 978-0-12-809451-8. 35. Dale, M.; Benson, S.M. Energy Balance of the Global Photovoltaic (PV) Industry—Is the PV Industry a Net Electricity Producer? Environ. Sci. Technol. 2013,47, 3482–3489. [CrossRef] 36. Price, L.; Kendall, A. Wind Power as a Case Study. J. Ind. Ecol. 2012,16, S22–S27. [CrossRef] 37. Cottrell, F. Energy and Society: The Relation between Energy, Social Change, and Economic Development; AuthorHouse: Bloomington, IN, USA, 2009; ISBN 978-1-4490-3169-5. 38. Hall, C.A.S.; Balogh, S.; Murphy, D.J.R. What is the Minimum EROI that a Sustainable Society Must Have? Energies 2009,2, 25–47. [CrossRef] 39. Hall, C.A.S.; Klitgaard, K. Energy and the Wealth of Nations: An Introduction to Biophysical Economics, 2nd ed.; Springer International Publishing: Berlin/Heidelberg, Germany, 2018; ISBN 978-3-319-66217-6. 40. White, L.A. Energy and the evolution of culture. Am. Anthropol. 1943,45, 335–356. [CrossRef] 41. Fizaine, F.; Court, V. Energy expenditure, economic growth, and the minimum EROI of society. Energy Policy 2016,95, 172–186. [CrossRef] 42. Lambert, J.; Hall, C.; Balogh, S.; Poisson, A.; Gupta, A. EROI of Global Energy Resources: Preliminary Status and Trends; College of Environmental Science and Forestry, State University of New York, United Kingdom Department for International Development: London, UK, 2012. 43. Raugei, M.; Leccisi, E. A comprehensive assessment of the energy performance of the full range of electricity generation technologies deployed in the United Kingdom. Energy Policy 2016,90, 46–59. [CrossRef] 44. Gagnon, N.; Hall, C.A.S.; Brinker, L. A Preliminary Investigation of Energy Return on Energy Investment for Global Oil and Gas Production. Energies 2009,2, 490–503. [CrossRef] 45. Masnadi, M.S.; Brandt, A.R. Energetic productivity dynamics of global super-giant oilfields. Energy Environ. Sci. 2017,10, 1493–1504. [CrossRef] 46. Smil, V. Energy Transitions: History, Requirements, Prospects; Praeger: Santa Barbara, CA, USA, 2010; ISBN 0-313-38177-1. 47. Arvesen, A.; Hertwich, E.G. Assessing the life cycle environmental impacts of wind power: A review of present knowledge and research needs. Renew. Sustain. Energy Rev. 2012,16, 5994–6006. [CrossRef] 48. Boccard, N. Capacity factor of wind power realized values vs. estimates. Energy Policy 2009 ,37, 2679–2688. [CrossRef] 49. Clack, C.T.M.; Qvist, S.A.; Apt, J.; Bazilian, M.; Brandt, A.R.; Caldeira, K.; Davis, S.J.; Diakov, V.; Handschy, M.A.; Hines, P.D.H.; et al. Evaluation of a proposal for reliable low-cost grid power with 100% wind, water, and solar. Proc. Natl. Acad. Sci. USA 2017,114, 6722–6727. [CrossRef] 50. Moriarty, P.; Honnery, D. Can renewable energy power the future? Energy Policy 2016,93, 3–7. [CrossRef] 51. Trainer, T. Some problems in storing renewable energy. Energy Policy 2017,110, 386–393. [CrossRef] 52. Trainer, T. Can Europe run on renewable energy? A negative case. Energy Policy 2013 ,63, 845–850. [CrossRef] 53. Trainer, T. A critique of Jacobson and Delucchi’s proposals for a world renewable energy supply. Energy Policy 2012,44, 476–481. [CrossRef] 54. Trainer, T. Can renewables etc. solve the greenhouse problem? The negative case. Energy Policy 2010 ,38, 4107–4114. [CrossRef] 55. Capell á n-P é rez, I.; de Blas, I.; Nieto, J.; de Castro, C.; Miguel, L.J.; Carpintero, Ó .; Mediavilla, M.; Lobej ó n, L.F.; Ferreras-Alonso, N.; Rodrigo, P.; et al. MEDEAS: A new modeling framework integrating global biophysical and socioeconomic constraints. Energy Environ. Sci. 2020,13, 986–1017. [CrossRef] 56. Capell á n-P é rez, I.; de Blas, I.; Nieto, J.; De Castro, C.; Miguel, L.J.; Mediavilla, M.; Carpintero, Ó .; Rodrigo, P.; Frechoso, F.; Cáceres, S. D4.1 MEDEAS Model and IOA Implementation at Global Geographical Level; MEDEAS Project: Barcelona, Spain, 2017. Energies 2020,13, 3036 38 of 42 57. Brandt, A.R. How Does Energy Resource Depletion Affect Prosperity? Mathematics of a Minimum Energy Return on Investment (EROI). Biophys. Econ. Resour. Qual. 2017,2, 2. [CrossRef] 58. Murphy, D.J.; Hall, C.A.S. Energy return on investment, peak oil, and the end of economic growth. Ann. N. Y. Acad. Sci. 2011,1219, 52–72. [CrossRef] 59. Capell á n-P é rez, I.; de Castro, C.; Arto, I. Assessing vulnerabilities and limits in the transition to renewable energies: Land requirements under 100% solar energy scenarios. Renew. Sustain. Energy Rev. 2017 ,77, 760–782. [CrossRef] 60. Moriarty, P.; Honnery, D. Ecosystem maintenance energy and the need for a green EROI. Energy Policy 2019 , 131, 229–234. [CrossRef] 61. UNEP. Recycling Rates of Metals. A Status Report; International Resource Panel; United Nations Environment Programme: Nairobi, Kenya, 2011. 62. Valero, A.; Ortego, A.; Calvo, G.; Valero, A.; C í rez, F.; Kimmich, C.; Cerny, M.; Kerschner, C.; Cernik, M.; Theofilidi, M.; et al. D2.1 Variables. Annex 9; CIRCE, MU, CRES & INSTM: Barcelona, Spain, 2016. 63. de Castro, C.; Mediavilla, M.; Miguel, L.J.; Frechoso, F. Global solar electric potential: A review of their technical and sustainable limits. Renew. Sustain. Energy Rev. 2013,28, 824–835. [CrossRef] 64. Frischknecht, R.; Itten, R.; Sinha, P.; de Wild-Scholten, M.; Zhang, J.; Fthenakis, V.; Kim, H.C.; Raugei, M.; Stucki, M. Life Cycle Inventories and Life Cycle Assessment of Photovoltaic Systems; IEA PVPS Task 12, Subtask 2.0; LCA. 2015. Available online: https://www.bnl.gov/pv/files/pdf/226_Task12_LifeCycle_ Inventories.pdf (accessed on 11 June 2020). 65. GWEC. GWEC Webpage; Global Wind Energy Council. Available online: https://gwec.net/global-figures/ wind-in-numbers/(accessed on 3 November 2019). 66. Muro Pereg, J.R.; Fern á ndez de la Hoz, J. ECOWIND. Life Cycle Assessment of 1 kWh Generated by a GAMESA Onshore Windfarm G90 2.O MW; GAMESA: Zamudio, Spain, 2013. 67. GWEC. Global Wind Report 2016; Global Wind Energy Council: Brussels, Belgium, 2017. 68. LondonArray. London Array. Available online: http://www.londonarray.com/(accessed on 28 March 2016). 69. SMart Wind. Hornsea Offshore Wind Farm Project One. Chapter 3: Project Description; Smart Wind Limited: London, UK, 2013. 70. Flury, K.; Frischknecht, R. Life Cycle Inventories of Hydroelectric Power Generation; ESU-Services, Fair Consulting in Sustainability: Schaffhausen, Switzerland, 2012; pp. 1–51. 71. Schellenberg, G.; Donnelly, C.R.; Holder, C.; Briand, M.-H.; Ahsan, R. Sedimentation, Dam Safety and Hydropower: Issues, Impacts and Solutions 2017. In Sedimentation and Hydropower: Impacts and Solutions. p. 24. Available online: https://www.hydroreview.com/wp-content/uploads/content/ dam/hydroworld/online-articles/2017/04/Sedimentation%20Dam%20Safety%20and%20Hydropower- %20Issues%20Impacts%20and%20Solutions.pdf (accessed on 11 June 2020). 72. IEA. IEA Sankey Webpage; International Energy Agency. Available online: https://www.iea.org/sankey/ (accessed on 1 November 2019). 73. Hertwich, E.G.; Gibon, T.; Bouman, E.A.; Arvesen, A.; Suh, S.; Heath, G.A.; Bergesen, J.D.; Ramirez, A.; Vega, M.I.; Shi, L. Integrated life-cycle assessment of electricity-supply scenarios confirms global environmental benefit of low-carbon technologies. Proc. Natl. Acad. Sci. USA 2015 ,112, 6277–6282. [CrossRef] 74. Kis, Z.; Pandya, N.; Koppelaar, R.H.E.M. Electricity generation technologies: Comparison of materials use, energy return on investment, jobs creation and CO 2 emissions reduction. Energy Policy 2018 ,120, 144–157. [CrossRef] 75. Schmied, M.; Knörr, W. Calculating GHG Emissions for Freight Forwarding and Logistics Services in Accordance with EN 16258; European Association for Forwarding; Transport, Logistics and Customs Services (CLECAT): Bruxelles, Belgium, 2012. 76. Hammond, G.; Jones, C. Inventory of Carbon & Energy (ICE) Version 2.0; Sustainable Energy Research Team (SERT) Department of Mechanical Engineering University of Bath: Bath, UK, 2011. 77. Rankin, J. Energy use in metal production. In Proceedings of the High Temperature Processing Symposium, Melbourne, Australia, 6–7 February 2012; pp. 6–7. 78. der Voet, E.V.; Oers, L.V.; Verboon, M.; Kuipers, K. Environmental Implications of Future Demand Scenarios for Metals: Methodology and Application to the Case of Seven Major Metals. J. Ind. Ecol. 2019 ,23, 141–155. [CrossRef] Energies 2020,13, 3036 39 of 42 79. Song, Y.S.; Youn, J.R.; Gutowski, T.G. Life cycle energy analysis of fiber-reinforced composites. Compos. Part A Appl. Sci. Manuf. 2009,40, 1257–1265. [CrossRef] 80. Guan,J.; Zhang,Z.; Chu,C.Quantificationofbuilding embodied energyin Chinausing an input–output-based hybrid LCA model. Energy Build. 2016,110, 443–452. [CrossRef] 81. Wu, K.-Y.; Wu, J.-H.; Huang, Y.-H.; Fu, S.-C.; Chen, C.-Y. Estimating direct and indirect rebound effects by supply-driven input-output model: A case study of Taiwan’s industry. Energy 2016 ,115, 904–913. [CrossRef] 82. Nieto, J.; Carpintero, Ó .; Miguel, L.J.; de Blas, I. Macroeconomic modelling under energy constraints: Global low carbon transition scenarios. Energy Policy 2019, 111090. [CrossRef] 83. Wiedmann, T.O.; Suh, S.; Feng, K.; Lenzen, M.; Acquaye, A.; Scott, K.; Barrett, J.R. Application of Hybrid Life Cycle Approaches to Emerging Energy Technologies—The Case of Wind Power in the UK. Environ. Sci. Technol. 2011,45, 5900–5907. [CrossRef] 84. European Commission. Critical Raw Materials for the UE. Report of the Ad-Hoc Working Group on Defining Critical Raw Materials; European Commission: Brussels, Belgium, 2010. 85. Elshkaki, A.; Graedel, T.E. Dynamic analysis of the global metals flows and stocks in electricity generation technologies. J. Clean. Prod. 2013,59, 260–273. [CrossRef] 86. Garc í a-Olivares, A.; Ballabrera-Poy, J.; Garc í a-Ladona, E.; Turiel, A. A global renewable mix with proven technologies and common materials. Energy Policy 2012,41, 561–574. [CrossRef] 87. Prior, T.; Giurco, D.; Mudd, G.; Mason, L.; Behrisch, J. Resource depletion, peak minerals and the implications for sustainable resource management. Glob. Environ. Chang. 2012,22, 577–587. [CrossRef] 88. Trainer, T. Estimating the EROI of whole systems for 100% renewable electricity supply capable of dealing with intermittency. Energy Policy 2018,119, 648–653. [CrossRef] 89. Pihl, E.; Kushnir, D.; Sand é n, B.; Johnsson, F. Material constraints for concentrating solar thermal power. Energy 2012,44, 944–954. [CrossRef] 90. Dahmus, J.B.; Gutowski, T.G. An environmental analysis of machining. In Proceedings of the ASME 2004 International Mechanical Engineering Congress and Exposition, Anaheim, CA, USA, 13–19 November 2004; American Society of Mechanical Engineers: New York, NY, USA, 2004; pp. 643–652. 91. Boustani, A.; Sahni, S.; Gutowski, T.; Graves, S. Appliance Remanufacturing and Energy Savings; Sloan School of Management: Cambridge, MA, USA, 2010. 92. Ciceri, N.D.; Gutowski, T.; Garetti, M. A tool to estimate materials and manufacturing energy for a product. In Proceedings of the 2010 IEEE International Symposium on Sustainable Systems and Technology, Arlington, VA, USA, 17–19 May 2010; IEEE: Piscataway, NJ, USA, 2010; pp. 1–6. 93. IRENA db. IRENA Resource; International Renewable Energy Agency. Available online: http://resourceirena. irena.org (accessed on 9 June 2020). 94. Dupont, E.; Koppelaar, R.; Jeanmart, H. Global available wind energy with physical and energy return on investment constraints. Appl. Energy 2018,209, 322–338. [CrossRef] 95. MTE. Estad í stica de la Industria de la Energ í a El é ctrica; Ministerio para la Transici ó n Ecol ó gica: Madrid, Spain, 2018. 96. MTE. Estad í stica de la Industria de la Energ í a El é ctrica; Ministerio para la Transici ó n Ecol ó gica: Madrid, Spain, 2017. 97. MTE. Estad í stica de la Industria de la Energ í a El é ctrica; Ministerio para la Transici ó n Ecol ó gica: Madrid, Spain, 2016. 98. MTE. Estad í stica de la Industria de la Energ í a El é ctrica; Ministerio para la Transici ó n Ecol ó gica: Madrid, Spain, 2015. 99. MTE. Estad í stica de la Industria de la Energ í a El é ctrica; Ministerio para la Transici ó n Ecol ó gica: Madrid, Spain, 2014. 100. Ferroni, F.; Guekos, A.; Hopkirk, R.J. Further considerations to: Energy Return on Energy Invested (ERoEI) for photovoltaic solar systems in regions of moderate insolation. Energy Policy 2017 ,107, 498–505. [CrossRef] 101. Farthing, A.; Carbajales-Dale, M.; Mason, S.; Carbajales-Dale, P.; Matta, P. Utility-Scale Solar PV in South Carolina: Analysis of Suitable Lands and Geographical Potential. Biophys. Econ. Resour. Qual. 2016 ,1, 8. [CrossRef] 102. WBGU. Future Bioenergy and Sustainable Land Use; German Advisory Council on Global Change (WBGU): Berlin, Germany, 2009; ISBN 978-1-84407-841-7. Energies 2020,13, 3036 40 of 42 103. Dupont, E.; Koppelaar, R.; Jeanmart, H. Global available solar energy under physical and energy return on investment constraints. Appl. Energy 2020,257, 113968. [CrossRef] 104. REN21. Renewables 2017. Global Status Report; REN 21: Paris, France, 2017. 105. Mart í n-Chivelet, N. Photovoltaic potential and land-use estimation methodology. Energy 2016 ,94, 233–242. [CrossRef] 106. Luque, A.; Hegedus, S. Handbook of Photovoltaic Science and Engineering; John Wiley & Sons: Chichester, UK, 2011. 107. Arto, I.; Capell á n-P é rez, I.; Lago, R.; Bueno, G.; Bermejo, R. The energy requirements of a developed world. Energy Sustain. Dev. 2016,33, 1–13. [CrossRef] 108. Dale, M.A.J. Global Energy Modelling: A Biophysical Approach (GEMBA). Ph.D. Thesis, University of Canterbury, Christchurch, New Zealand, 2010. 109. Schoenberg, B.; Hall, C.A.S. The Energy Return of (Industrial) Solar—Passive Solar, PV, Wind and Hydro (#5 of 6). The Oil Drum 2008. Available online: http://theoildrum.com/node/3910 (accessed on 10 June 2020). 110. Pillai, U. Drivers of cost reduction in solar photovoltaics. Energy Econ. 2015,50, 286–293. [CrossRef] 111. Dale, M.; Krumdieck, S.; Bodger, P. A Dynamic Function for Energy Return on Investment. Sustainability 2011,3, 1972–1985. [CrossRef] 112. Van de Ven, D.-J.; Capell á n-P é rez, I.; Arto, I.; Cazcarro, I.; De Castro, C.; Patel, P.; Gonz á lez-Eguino, M. The potential land use requirements and related land use change emissions of solar energy. Nat. Sustain. 2020. under review. 113. Calvo, G.; Mudd, G.; Valero, A.; Valero, A. Decreasing Ore Grades in Global Metallic Mining: A Theoretical Issue or a Global Reality? Resources 2016,5, 36. [CrossRef] 114. Harmsen, J.H.M.; Roes, A.L.; Patel, M.K. The impact of copper scarcity on the efficiency of 2050 global renewable energy scenarios. Energy 2013,50, 62–73. [CrossRef] 115. Mudd, G.M. The Environmental sustainability of mining in Australia: Key mega-trends and looming constraints. Resour. Policy 2010,35, 98–115. [CrossRef] 116. Steffen, B.; Hischier, D.; Schmidt, T.S. Historical and projected improvements in net energy performance of power generation technologies. Energy Environ. Sci. 2018,11, 3524–3530. [CrossRef] 117. Louwen, A.; van Sark, W.G.J.H.M.; Faaij, A.P.C.; Schropp, R.E.I. Re-assessment of net energy production and greenhouse gas emissions avoidance after 40 years of photovoltaics development. Nat. Commun. 2016 ,7, 1–9. [CrossRef] 118. Sers, M.R.; Victor, P.A. The Energy-missions Trap. Ecol. Econ. 2018,151, 10–21. [CrossRef] 119. Heun, M.K.; de Wit, M. Energy return on (energy) invested (EROI), oil prices, and energy transitions. Energy Policy 2012,40, 147–158. [CrossRef] 120. Parrique, T.; Barth, J.; Briens, F.; Kerschner, C.; Kraus-Polk, A.; Kuokkanen, A.; Spangenberg, J.H. Decoupling Debunked—Evidence and Arguments against Green Growth as a Sole Strategy for Sustainability; European Environmental Bureau (EEB): Brussels, Belgium, 2019. 121. Dale, M.; Krumdieck, S.; Bodger, P. Global energy modelling—A biophysical approach (GEMBA) Part 2: Methodology. Ecol. Econ. 2012,73, 158–167. [CrossRef] 122. Sgouridis, S.; Csala, D.; Bardi, U. The sower’s way: Quantifying the narrowing net-energy pathways to a global energy transition. Environ. Res. Lett. 2016,11, 094009. [CrossRef] 123. European Commission. A Roadmap for Moving to a Competitive Low Carbon Economy in 2050; Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions; European Commission: Brussels, Belgium, 2011. 124. Jacobs, M. Green growth: Economic theory and political discourse. Centre for Climate Change Economics and Policy Working Paper No. 108; Grantham Research Institute on Climate Change and the Environment Working Paper No. 92. 2012. Available online: http://www.lse.ac.uk/GranthamInstitute/publication/greengrowth-economic-theory-and-political-discourse-working-paper-92/(accessed on 11 June 2020). 125. OECD. OECD Work on Green Growth. Available online: http://www.oecd.org/greengrowth/ oecdworkongreengrowth.htm (accessed on 12 March 2018). 126. OECD. Towards Green Growth; Organisation for Economic Co-operation and Development: Paris, France, 2011. 127. UNEP. Towards a Green Economy: Pathways to Sustainable Development and Poverty Eradication; United Nations Environment Programme: Nairobi, Kenya, 2011. Energies 2020,13, 3036 41 of 42 128. World Bank Inclusive Green Growth: The Pathway to Sustainable Development; World Bank Publications: Washington, DC, USA, 2012; ISBN 978-0-8213-9551-6. 129. Demaria, F.; Schneider, F.; Sekulova, F.; Martinez-Alier, J. What is Degrowth? From an Activist Slogan to a Social Movement. Environ. Values 2013,22, 191–215. [CrossRef] 130. IEA;IRENA.Perspectives for the Energy Transition. Investment Needs for a Low-Carbon Energy System; International Energy Agency: Paris, France; International Renewable Energy Agency: Abu Dhabi, UAE, 2017. 131. Nemet, G.F. Beyond the learning curve: Factors influencing cost reductions in photovoltaics. Energy Policy 2006,34, 3218–3232. [CrossRef] 132. Lawn, P. How well are resource prices likely to serve as indicators of natural resource scarcity? Int. J. Sustain. Dev. 2004,7, 369–397. [CrossRef] 133. Norgaard, R.B. Economic indicators of resource scarcity: A critical essay. J. Environ. Econ. Manag. 1990 ,19, 19–25. [CrossRef] 134. Reynolds, D.B. The mineral economy: How prices and costs can falsely signal decreasing scarcity. Ecol. Econ. 1999,31, 155–166. [CrossRef] 135. Rubin, E.S.; Azevedo, I.M.L.; Jaramillo, P.; Yeh, S. A review of learning rates for electricity supply technologies. Energy Policy 2015,86, 198–218. [CrossRef] 136. Daly, H.E. Ecological Economics and Sustainable Development; Edward Elgar Publishing: Cheltenham, UK, 2007. 137. Smil, V. Power Density: A Key to Understanding Energy Sources and Uses; The MIT Press: Cambridge, MA, USA, 2015; ISBN 978-0-262-02914-8. 138. Gernaat, D.E.H.J.; Bogaart, P.W.; van Vuuren, D.P.; Biemans, H.; Niessink, R. High-resolution assessment of global technical and economic hydropower potential. Nat. Energy 2017,2, 821. [CrossRef] 139. Connolly, D.; Lund, H.; Mathiesen, B.V. Smart Energy Europe: The technical and economic impact of one potential 100% renewable energy scenario for the European Union. Renew. Sustain. Energy Rev. 2016 ,60, 1634–1653. [CrossRef] 140. Moeller, D.; Murphy, D. Net Energy Analysis of Gas Production from the Marcellus Shale. Biophys. Econ. Resour. Qual. 2016,1, 5. [CrossRef] 141. Murphy, D.J. The implications of the declining energy return on investment of oil production. Philos. Trans. R. Soc. A 2014,372, 20130126. [CrossRef] [PubMed] 142. Ramorakane, R.J.; Dinter, F. Evaluation of parasitic consumption for a CSP plant. AIP Conf. Proc. 2016 ,1734, 070027. [CrossRef] 143. García-Olivares, A. Energy for a sustainable post-carbon society. Sci. Mar. 2016,80, 257–268. [CrossRef] 144. Latunussa, C.E.L.; Ardente, F.; Blengini, G.A.; Mancini, L. Life Cycle Assessment of an innovative recycling process for crystalline silicon photovoltaic panels. Sol. Energy Mater. Sol. Cells 2016 ,156, 101–111. [CrossRef] 145. Martinez-Alier, J. The Environmentalism of the Poor: A Study of Ecological Conflicts and Valuation; Edward Elgar Publishing: Cheltenham, UK, 2003. 146. UNEP Environmental Risks and Challenges of Anthropogenic Metals Flows and Cycles; International Resource Panel; United Nations Environment Programme: Cheltenham, UK, 2013. 147. Haapala, K.R.; Prempreeda, P. Comparative life cycle assessment of 2.0 MW wind turbines. Int. J. Sustain. Manuf. 2014,3, 170–185. [CrossRef] 148. Ribrant, J.; Bertling, L. Survey of failures in wind power systems with focus on Swedish wind power plants during 1997-2005. In Proceedings of the 2007 IEEE Power Engineering Society General Meeting, Tampa, FL, USA, 24–28 June 2007; pp. 1–8. 149. EUROELECTRIC. Power Distribution in Europe. Facts & Figures; Eurelectric: Brussels, Belgium, 2013. 150. NREL. Renewable Electricity Futures Study (Entire Report); National Renewable Energy Laboratory: Golden, CO, USA, 2012. 151. Bumby, S.; Druzhinina, E.; Feraldi, R.; Werthmann, D.; Geyer, R.; Sahl, J. Life Cycle Assessment of Overhead and Underground Primary Power Distribution. Environ. Sci. Technol. 2010,44, 5587–5593. [CrossRef] 152. Jorge, R.S.; Hawkins, T.R.; Hertwich, E.G. Life cycle assessment of electricity transmission and distribution—Part 1: Power lines and cables. Int. J. Life Cycle Assess. 2012,17, 9–15. [CrossRef] 153. Jorge, R.S.; Hawkins, T.R.; Hertwich, E.G. Life cycle assessment of electricity transmission and distribution—Part 2: Transformers and substation equipment. Int. J. Life Cycle Assess. 2012 ,17, 184–191. [CrossRef]