The Paris Agreement's 1.5°C warming target necessitates a large-scale transition to low-carbon energy. However, this transition itself demands significant energy and resources, raising concerns about potential energy scarcity and increased emissions. Previous research hints that rapid low-carbon infrastructure growth could consume a considerable portion of global energy supply, and because current energy production relies heavily on fossil fuels, the transition process might become a significant emissions source. While the IPCC's Special Report on Global Warming of 1.5 °C outlines various low-carbon energy pathways, it lacks estimates of the energy and emissions associated with building and maintaining a global low-carbon energy system. This knowledge gap is crucial, as studies suggest renewables may have a lower energy return on energy invested (EROI) compared to fossil fuels, potentially leading to decreased net energy available to society and creating an "energy-emissions trap". However, recent research indicates that this hypothesis might be overstated due to potential overestimation of fossil fuel EROI and underestimation of renewable EROI improvements. Life-cycle assessments (LCAs) have been used to quantify climate change impacts of energy technologies, but typically only for present-day technologies in specific case studies, lacking dynamic analyses of entire global systems. Existing studies on the energy transition's energy and emissions implications are limited in scope, often focusing solely on electricity generation or a single low-carbon pathway. This research addresses these gaps by comprehensively estimating the energy and emissions associated with the global energy system during a low-carbon transition.

Several studies have explored the energy requirements and emissions implications of a low-carbon energy transition. Some suggest that renewables exhibit lower EROI compared to existing energy systems, potentially leading to reduced net energy availability for society. This could result in an "energy-emissions trap", where ambitious climate mitigation efforts cause both reduced energy availability and significant carbon emissions. Conversely, other studies challenge this notion, suggesting that overestimations of fossil fuel EROI and underestimations of renewable EROI improvements may have exaggerated the potential for lower energy availability. Life cycle assessment (LCA) is another method used to analyze the climate impacts of different energy technologies. However, most LCAs focus on existing technologies and case studies, lacking dynamic analyses that track changing technological impacts over time or assess the cumulative impacts of decarbonizing the entire global energy system. Notable exceptions include Pehl et al.'s study estimating 82 GtCO2eq of cumulative emissions from power plant construction, operation and maintenance from 2010–2050, and Di Felice et al.'s assessment of indirect emissions from the EU's renewable energy strategy, calculating 25 GtCO2eq from 2020–2050. These studies, however, only consider electricity generation and specific low-carbon pathways.

This study estimates the energy required and carbon emissions involved in constructing, operating, and maintaining the global energy system throughout a low-carbon energy transition. It distinguishes between energy and emissions related to the energy system itself and those available for other societal needs. The analysis utilizes a consumption-based accounting approach, employing EROI analysis at the final energy stage to estimate direct and indirect energy use and emissions associated with constructing, operating, and maintaining the energy system and supplying energy to society. The research integrates EROI estimates of current energy technologies with projections of future EROI changes due to technological improvements estimated using energetic experience curves. To account for the range of present-day estimates, uncertainties in technological change, and resource availability, the study reports estimates using low, median, and high-EROI values, representing the first, second, and third quartiles of the interquartile range. A consistent energy system boundary is applied to all energy technologies, extending from extraction to point-of-use, addressing inconsistencies in EROI values across studies and carriers. Three EROI scenarios are developed to account for variations in renewable EROI growth and bioenergy EROI. The study examines energy system emissions across fourteen 1.5°C-compatible mitigation pathways from six Integrated Assessment Models (IAMs), including four illustrative pathways from the IPCC's Special Report on Global Warming of 1.5 °C (LED, S1-A, S2-M, and S5-R) and ten additional pathways. Energy system emissions are calculated as the product of energy for the energy system and its carbon intensity, differentiating between four energy carriers (electricity, gases, liquid fuels, and solids) and three life-cycle stages (construction, operation and maintenance, and decommissioning). The analysis also considers changes in carbon intensity over time, accounting for the declining share of conventional fossil fuels. A panel data analysis is used to identify factors influencing energy system emissions, controlling for heterogeneity across pathways.

The study reveals substantial cumulative carbon emissions associated with the energy system during the transition, representing a considerable share of total cumulative emissions under various 1.5°C-compatible scenarios. The fourteen-pathway average is 195 GtCO2 for the median-EROI scenario, ranging from 185 GtCO2 (high-EROI) to 290 GtCO2 (low-EROI). This represents an average of 21% of total emissions (median-EROI), 20% (high-EROI), and 31% (low-EROI). Energy system emissions increase over time, consuming a growing share of total emissions. The share is estimated to increase 2-5 times its current value by 2060, depending on EROI assumptions. After 2060, this share stabilizes in most pathways as they achieve high decarbonization. Pathways with slower decarbonization, higher energy use, and greater BECCS deployment exhibit higher energy system emissions. The fourteen-pathway average of energy system emissions increases from 10% (median-EROI) in 2006–2015 to 27% in 2050 and 40% by 2100. A high share of energy system emissions, particularly in the low-EROI scenario, may constrain residual emissions for other sectors (aviation, steel and cement production). In some pathways, the energy system may consume all residual emissions by 2080 under all EROI scenarios. The initial energy required to build a low-carbon energy system causes only a small increase in annual emissions, most notably in pathways with higher energy use and fossil fuel reliance beyond 2030. In rapid decarbonization pathways, this increase is minimal, with pathways prioritizing renewables and nuclear over bioenergy (especially BECCS) achieving lower cumulative energy system emissions. BECCS, due to its low EROI and energy conversion efficiency, exhibits higher emissions per unit of energy generated. The study also indicates a decline in net energy available to society during the transition, ranging from 10% to 34% depending on the mitigation pathway. In pathways with lower energy use and rapid decarbonization, this decline is most significant during the initial construction phase. However, this does not necessarily imply energy scarcity. The EROI of the overall energy system depends on technology choice, declining in pathways that prioritize bioenergy and fossil fuels with carbon capture and storage, and increasing in pathways focusing on renewable energy technologies. The panel data analysis reveals that energy system emissions are driven by final energy use and the choice of energy technologies, while the overall EROI of the energy system plays a smaller role, particularly after 2040.

The findings highlight the substantial and growing contribution of energy system emissions to total emissions during a low-carbon transition. The estimated 195 GtCO2 from energy system emissions implies around 0.1 °C of additional warming, which, while significant, is small compared to long-term emissions reductions from decarbonization. The study does not find evidence of a major short-term emissions increase from intensified decarbonization efforts. Rather, slower decarbonization pathways with continued fossil fuel use and greater reliance on negative emissions technologies have higher emissions. This underscores the importance of rapid decarbonization and the limitations of negative emissions technologies, which appear less efficient in mitigating emissions than assumed in many models. The reduction in net energy availability during the transition, though substantial, doesn't necessitate energy scarcity. Efficient energy use, consumption shifts, and technological advancements can offset reduced net energy while maintaining access to essential energy services. The study's results align with prior research on net energy reduction, but show that these reductions occur earlier, within the first decade of the transition. The observed differences in energy system emissions and net energy reduction across various pathways highlight the importance of technology choice and energy demand management. The greater reliance on direct fossil fuel substitutes in certain pathways leads to higher emissions and less net energy. However, a comprehensive discussion about the relative merits of different decarbonization strategies is beyond this study's scope.

This research demonstrates the significance of explicitly modeling energy system emissions and energy requirements in low-carbon transition pathways. The considerable and growing share of emissions from the energy system itself needs to be considered when setting targets and designing policies. While the transition may temporarily decrease net energy per capita, the long-term gains from decarbonization far outweigh the short-term costs. Future research should incorporate the energy requirements and emissions associated with consumption-end infrastructure (e.g., electric vehicles, charging stations) and explore the use of alternative energy modeling approaches beyond traditional IAMs, which often favor certain technologies and may underestimate the potential of renewables. Integrating dynamic EROI analysis into IAMs could lead to more accurate projections of energy system dynamics and inform more efficient decarbonization strategies.

The study uses globally averaged EROI values, neglecting potential regional variations in energy production and transformation processes. The estimations of energy system emissions might underestimate the actual emissions in pathways heavily relying on bioenergy due to limitations in the availability of data on total positive emissions from BECCS technologies. The analysis is based on existing IAM scenarios and may not capture all possible future energy transition pathways. Finally, the study focuses on the energy system; incorporating the emissions associated with consumption-end infrastructure transformation requires future investigation.