Limiting global warming to 1.5°C or 2°C necessitates achieving net-zero or net-negative CO₂ emissions by the end of the century. This requires a significant transformation of energy systems, involving fossil fuel reduction, efficiency improvements, electrification, and carbon management. While numerous studies have assessed net-zero pathways, a systematic evaluation of common features and trade-offs across global scenarios at the net-zero point is lacking. This study addresses this gap by analyzing a large ensemble of Integrated Assessment Model (IAM) scenarios to identify robust features of net-zero energy systems and their regional characteristics, considering the implications for current energy policies and sustainable development goals. The study focuses on the characteristics of energy systems at the point of net-zero CO2 emissions to inform policy discussions and to explore approaches not fully represented in existing scenarios. The inertia of existing fossil fuel infrastructure and the need to reconcile energy system transformation with sustainable development goals add complexity to this challenge. The common features and trade-offs of scenarios at net-zero can potentially inform various policy approaches, including those not yet represented in current scenario pathways.

Existing literature includes detailed analyses of specific energy services and technologies, and some studies have examined mitigation pathways in IAM scenarios limiting warming to below 1.5°C. However, a systematic assessment of common features and trade-offs of these scenarios at the point of net-zero CO₂ emissions has not been conducted. This research builds upon previous work by focusing specifically on the characteristics of energy systems at the moment global CO₂ emissions reach net-zero. This allows for a unique perspective on policy implications and the exploration of potential approaches to reaching net-zero not fully captured in existing scenarios.

The study analyzes 177 IAM scenarios from the public 1.5°C Scenario Database (SR1.5 database), where global CO₂ emissions reach net-zero by 2100. The scenarios cover various levels of projected global warming (below 1.5°C, below 2°C, and above 2°C), including overshoot scenarios. The analysis assesses global and regional energy use, energy sources (renewable and non-renewable), residual emissions, electrification levels, and climate policies (represented by carbon prices). The data processing involved interpolation of annual data using second-order polynomials to address the 5 or 10-year timesteps of the original data. Only CO₂ emissions were considered, excluding CH₄ and N₂O, due to the focus on current net-zero CO₂ policies, the lack of practical pathways to net-zero for these other gases, and the fact that N₂O emissions are primarily related to agriculture. The scenarios were categorized into six regions (global and five world regions defined in the SR1.5 database) and three warming levels. Data inconsistencies across IAMs necessitated the selection of a subset of 177 scenarios containing all required output variables. The authors used one-way ANOVA and correlation analysis to examine relationships between various variables. The analysis uses seven key output variables: CO₂ emissions (total, net energy and industrial processes, and net AFOLU), population, GDP (PPP), primary energy (direct equivalent), carbon sequestration from BECCS, carbon price, and final energy consumption (total and electricity share). Residual CO₂ emissions were calculated by combining residual emissions from energy and industrial processes with BECCS sequestration (which offsets residual emissions). The analysis was performed using JupyterLab.

Across the 177 net-zero scenarios, renewable energy sources account for approximately 60% of primary energy, with biomass constituting a significant portion. Electricity represents about half of final energy consumption. Scenarios aiming for less warming (below 1.5°C) generally exhibit lower energy use and GDP per capita than those allowing for more warming (below 2°C or above 2°C). Warmer scenarios tend to achieve net-zero emissions later, and higher warming is strongly correlated with later net-zero years (r=0.73). The timing of peak emissions, however, is fairly consistent across scenarios, mostly around 2020. The share of renewable primary energy and the share of non-biomass renewables remain relatively consistent across warming levels, though they vary considerably across individual scenarios. The median share of primary energy from fossil fuels is 33%, ranging from 3-64% across scenarios. Residual emissions intensity remains substantial in many scenarios. Electrification levels are similar across warming levels, with only subtle differences; however, final energy use per capita varies substantially across warming levels. The level of electrification is positively correlated with the net-zero year, suggesting later scenarios compensate with higher electrification levels due to increased time for end-use transitions. Carbon prices show a wide range without a clear relationship to warming levels or BECCS deployment. The magnitude of carbon sequestration from BECCS increases with higher warming levels. Regional differences are substantial; Asia and OECD/EU regions typically have the highest energy consumption. Latin America shows a higher share of renewable energy and a higher proportion of biomass. Regional variations in electrification are smaller than those in energy consumption per capita. Residual emissions and negative emissions (BECCS) are unevenly distributed across regions, with some regions consistently net-positive while others are net-negative. This uneven distribution has implications for burden-sharing and equity. In warmer scenarios, net emissions from agriculture and land use are less negative, residual emissions are higher, and larger negative emissions from BECCS are needed to achieve net-zero. Analysis of relationships between scenario characteristics reveals that scenarios with higher electrification levels tend to have greater shares of renewable and non-biomass renewable energy but less energy conservation. Scenarios with greater energy conservation tend to have lower shares of non-biomass renewables. Relationships between negative emissions and other parameters are less clear and likely influenced by IAM structures and assumptions. Transportation was the main source of residual emissions in most scenarios where sector-level detail was available. Table 1 provides a statistical summary of these findings.

The findings challenge the notion of "electrifying everything" as the sole solution for decarbonization. While electrification is important, a significant portion of final energy continues to come from solid, liquid, and gaseous fuels even in net-zero scenarios. This highlights the need for decarbonizing hard-to-electrify sectors. The substantial residual emissions in many scenarios underscore the critical role of negative emissions technologies. However, the uneven regional distribution of residual and negative emissions poses challenges for equitable burden-sharing and necessitates policy consideration. The differences in energy use, especially in 1.5°C scenarios, suggest the importance of reducing energy demand. The observation that later, less ambitious scenarios exhibit slightly higher electrification could result from greater transition time. Overall, the results emphasize the need for policies that reduce reliance on negative emissions and focus on the entire energy system, including hard-to-electrify sectors, considering regional equity and sustainability.

This study provides valuable insights into the common characteristics and regional variations of net-zero energy systems. The findings highlight the importance of renewable energy, but also underscore the continued role of fossil fuels, the necessity of negative emissions technologies, and the challenges of equitable burden-sharing. Future research should investigate more detailed regional and technological aspects of net-zero pathways, considering the potential impact of rapid technological change and addressing the uncertainties surrounding negative emissions technologies at large scales. Refining the models to incorporate more detailed representations of agriculture, forestry, and other land use (AFOLU) sectors and non-CO₂ emissions is essential for a more holistic understanding of net-zero emissions pathways.

The study uses IAM scenarios which, while useful for large-scale analysis, have limitations. They may underestimate the role of variable renewables due to lower technological, temporal, and spatial resolution. The exclusion of CH₄ and N₂O emissions limits the comprehensiveness of the analysis and the fact that the study does not explicitly consider the details of agriculture, forestry, and other land use (AFOLU) sectors and non-CO₂ emissions; however, these aspects are accounted for in the IAM frameworks themselves, which consistently include the linkages and tradeoffs between AFOLU and non-CO₂ emissions. Model differences and scenario constraints affect the range of outputs, influencing the interpretation of relationships between variables. The study's focus on the net-zero year rather than pathways to net-zero may affect the generalizability of the results. The limited number of scenarios, due to data availability issues, reduces statistical power. Finally, IAMs do not fully represent societal dynamics and political economic factors that drive national emissions reduction strategies.