The global transition to net-zero carbon emissions by 2050 necessitates a significant shift away from fossil fuels, as highlighted by the Glasgow Climate Pact. However, the feasibility of various decarbonization pathways is subject to several constraints, including biophysical, economic, political, and technological limitations. Existing studies on energy return on investment (EROI) – the ratio of energy output to energy invested – have yielded diverse results due to methodological inconsistencies, varying boundary conditions, and the inclusion or exclusion of enabling technologies. This study addresses these shortcomings by employing a newly developed systemwide EROI model (LUT-EROI) to evaluate the sustainability risk of nine global energy transition scenarios. This holistic approach incorporates the optimal interoperability of multiple processes, addressing fundamental gaps in existing EROI estimation techniques. The study analyzes the impact of electricity generation mix on EROI and investigates its relationships with levelized cost of electricity (LCOE) and CO2 emissions, enhancing our understanding of pathway selection for an optimal system development strategy.

Prior research on EROI has been marked by methodological inconsistencies, including diverse concepts of EROI, varying boundary conditions, and differing approaches to comparing fossil fuel and renewable energy technologies (with or without enabling technologies). While some studies have questioned the plausibility of 100% renewable energy systems based on net energy production, recent research employing advanced systemwide EROI approaches has challenged these conclusions, highlighting the importance of capturing the impact of optimal interoperability of multi-processes and addressing typical methodological gaps in EROI estimations. This study builds upon this recent work to provide a more comprehensive and robust analysis of global energy transition scenarios.

The study utilizes the newly developed Excel-based LUT-EROI model, which improves upon existing EROI studies by implementing a holistic approach for estimating primary energy quality at the electricity level. It leverages a broader cumulative energy demand (CED) database for technologies based on life cycle assessment (LCA) databases, integrating EROI estimation with energy system model output. The analysis focuses on nine global energy transition scenarios presented in Aghahosseini et al. These scenarios are categorized into two groups: (i) optimizations of the Best Policy Scenario (BPS) and faster transition scenarios based on BPS, and (ii) replications of scenarios from the International Energy Agency (IEA) and Teske and the German Aerospace Centre (DLR). The scenarios differ in their decarbonization targets (achieving net-zero CO2 emissions before 2050, by 2050, or after 2050), energy mixes, technology diversity, and transition speeds. The LUT-EROI model estimates CED values following LCA rules, maintaining consistent primary energy quality. It incorporates a mathematical algorithm capturing changes in technology operation rules implemented in the LUT Energy System Transition Model (LUT-ESTM), a cost-optimization energy system model used for analyzing short- and medium-term goals to achieve 100% net-zero CO2 power systems. The model considers various factors such as the electricity generation mix, storage capacity, curtailment, and the embodied energy requirements for natural gas and oil sources. The analysis considers the point of electricity generation and consumption, with estimates at the point of final energy consumption presented in the main text and generation-level estimations (excluding transmission and distribution losses) provided in the supplementary information. A sensitivity analysis is conducted to assess the impact of variations in the embodied energy requirements for natural gas and oil on the EROI trends. The model also accounts for technology evolution by incorporating energy learning rates for PV and battery technologies and considering projected capacities and market shares for other technologies. Annual energy investment flows (AEF) are estimated to further analyze the sustainability risk of the systemwide EROI.

The study found that all nine scenarios resulted in systemwide EROI values above 16, remaining consistently above the net energy cliff (around 10) throughout the 30-year transition period. However, EROI trends exhibited significant variations depending on the specific transition pathway. The LUT-BPS scenarios showed an initial increase in EROI followed by a continuous decline, attributed to the increasing share of renewable energy sources and the need for enabling technologies (mainly batteries and gas storage). The rate of decline differed across scenarios, with faster transition scenarios showing sharper declines initially but stabilizing later. The IEA scenarios displayed an initial drop in EROI followed by a slight rebound, primarily due to the slower increase in variable renewable energy (VRE) penetration and the continued utilization of nuclear power and natural gas. The Teske/DLR scenarios showed similar characteristics to IEA scenarios until 2040, after which a decline emerged, influenced by the capacity expansion of geothermal and concentrating solar power plants. The analysis revealed a strong correlation between EROI and VRE penetration; EROI initially increased with VRE penetration up to 50%, after which it declined as VRE penetration increased further, necessitating greater storage capacity. The study also demonstrated a relationship between EROI and LCOE, finding that low-cost solutions correlated with low EROI. While low-EROI systems were found to be cleaner options, higher EROI systems tended to have higher CO2 emissions. A sensitivity analysis revealed that IEA scenarios were highly sensitive to changes in the embodied energy requirements of natural gas and oil, while BPS scenarios were relatively insensitive due to their high reliance on renewable energy. Regional analyses showed significant variations in EROI trends due to uneven resource distributions and system compositions. Annual energy investment flows remained below 16% of final energy consumption in all scenarios, suggesting that none of the scenarios pose significant sustainability risks from this perspective.

The findings of this study underscore the importance of considering systemwide EROI in the planning and design of sustainable power systems. The transition to 100% renewable energy does not necessarily lead to a collapse in net energy, as indicated by the EROI values consistently exceeding the net energy cliff. However, the study highlights the trade-offs between rapid decarbonization, EROI, and the cost of enabling technologies. Faster transition scenarios, while achieving lower CO2 emissions, may experience sharper declines in EROI in the short term due to significant investments in storage and other enabling technologies. The study emphasizes the significance of technological diversification, strategic technology selection, and optimization of system design and operation to mitigate the negative impacts on EROI. The sensitivity of some scenarios to fossil fuel resource availability further highlights the need for robust energy policy and transition planning. Future studies should investigate sector coupling impacts on EROI trends and explore the economic impacts of varying EROI levels on society.

This study demonstrates the feasibility of various net-zero CO2 power system scenarios from a physical EROI perspective. However, the optimal pathway depends significantly on the balance between the speed of transition and the associated decline in EROI, which is largely impacted by the choice and deployment of enabling technologies. The results support the technological optimism surrounding 100% renewable energy systems. Further research should focus on the integration of socio-economic and political factors into EROI analysis and investigate the impacts of sector coupling on systemwide EROI trends.

The study's limitations include the use of aggregated data, which may obscure regional variations and specific local challenges. The reliance on a single CED database and the linear extrapolation method for estimating CED values for technologies may introduce uncertainties. Unforeseen technological advancements and the exclusion of recycling of materials and waste in the analysis might also affect the findings. The impact of transmission and distribution networks was not fully incorporated due to the limited data available. Furthermore, the study does not consider the intersectional impacts of socio-economic-political variations or the direct and indirect effects of climate change on power systems.