Energy and socioeconomic system transformation through a decade of IPCC-assessed scenarios

Nature Climate Change | Volume 15 | February 2025 | 218–226 218 nature climate change Analysis https://doi.org/10.1038/s41558-024-02198-6 Energy and socioeconomic system transformation through a decade of IPCC-assessed scenarios D. J. van de Ven 1 , S. Mittal 2,3, A. Nikas 4, G. Xexakis 5, A. Gambhir 2, L. Hermwille 6, P. Fragkos 7, W. Obergassel 6, M. Gonzalez-Eguino 1,8,9, F. Filippidou7, I. Sognnaes 3, L. Clarke10 & G. P. Peters 3 Charting future emissions pathways is a central tenet of IPCC assessment reports (AR), yet it is unclear how underlying drivers (including around policy and technology) have influenced the evolution of emissions pathways. Here we compare scenarios in AR5 and AR6 and find that scenarios without specific climate policies enforced have shifted lower in each scenario generation, owing to falling low-carbon technology costs and reduced expectations for economic growth, reducing fossil-fuel shares in energy and industry. Mitigation pathways consistent with 1.5–2 °C have seen increasing electrification rates and higher shares of variable renewables in electricity in more recent scenario generations, implying reduced reliance on coal, nuclear, bioenergy and carbon capture and storage, reflecting changing costs. Despite the shrinking carbon budget due to insufficient recent climate action, mitigation costs have not increased given more optimistic low-carbon technology cost projections. Moving forward, scenario producers must continually recalibrate to keep abreast of technology, policy and societal developments to remain policy relevant. The decade in between the Sixth (AR6)1 and Fifth (AR5)2 assessment reports of the IPCC has seen dramatic cost reductions and performance improvements in key clean technologies such as solar photovoltaic (PV), battery storage and wind power3–7. However, it has also seen relative stagnation in long-considered mitigation ‘game-changers’ 8 , resulting in concerns over the plausibility of the diffusion scales assumed in modelled mitigation pathways—regarding carbon capture and storage (CCS) 9 or carbon dioxide removal (CDR), particularly bioenergy with CCS (BECCS) 10–13 . The same decade saw the introduction of low-carbon policies and targets14 that have helped shift away from ‘baseline’ (no-climate-policy) scenarios 15 —alongside increased climate awareness among citizens 16 and disruptive socioeconomic events amidst lingering effects of the Great Recession that include a pandemic, international conflict and large-scale energy crisis. These developments have had varied implications for climate action, thereby altogether offering useful opportunities to assess the capacity of integrated assessment models (IAMs) to keep up17 and to understand how the insights emerging from IAM exercises have changed throughout the decade. In this period, the scientific field of IAMs has grown substantially in research output18 and user interest 19 , driven by increasing dedication to the climate emergency 20 and improvements in computational power 21 . This growth has allowed thousands of scenarios projecting mitigation pathways with predefined emissions or climate targets to be published in the literature, revealing insights on how emission reductions might be prioritized across countries and sectors. To fully reflect on what can be learnt from recent scenario literature and understand how potential developments onwards can Received: 20 February 2024 Accepted: 31 October 2024 Published online: 3 January 2025 Check for updates 1Basque Centre for Climate Change, Leioa, Spain. 2Grantham Institute, Imperial College, London, UK. 3CICERO Center for International Climate Research, Oslo, Norway. 4National Technical University of Athens, Athens, Greece. 5HOLISTIC, Athens, Greece. 6Wuppertal Institute for Climate, Environment and Energy, Wuppertal, Germany. 7E3 Modelling, Athens, Greece. 8IKERBASQUE, Basque Foundation for Science, Bilbao, Spain. 9University of the Basque Country (UPV/EHU), Bilbao, Spain. 10Bezos Earth Fund, Washington, DC, USA. e-mail: [email protected] Nature Climate Change | Volume 15 | February 2025 | 218–226 219 Analysis https://doi.org/10.1038/s41558-024-02198-6 for systematic comparison of several dimensions of mitigation. We additionally include inputs from the Special Report on Emissions Scenarios (SRES), which was used to inform the AR3 and AR4 (ref. 27). As baseline-only scenarios, SRES are compared exclusively to baseline scenarios across ensembles. All observations throughout this section are ‘model means’—that is, the mean indicator value from all scenarios of similar climate ambition from the same model (Methods), guaranteeing results are not biased towards specific models with a large number of scenarios in IPCC assessments. Evolution of baseline scenarios Baseline scenarios are an important reference point. First, they are unconstrained by carbon budgets and mostly driven by assumed continuation of current technology, economic and societal trends. Analysing baseline scenarios, therefore, is important to understand the significance of the observed differences in more ambitious scenarios. Second, the transformative potential of a project or political instrument has been argued to correspond to its potential to shift from an assumed business-as-usual (BAU)28. BAU scenarios should reflect the expected change under socioeconomic assumptions, be addressed 22 , it is instructive to assess what scenario database ensembles say about the evolving role of technology and climate policy 23 . Here, we present such analysis comparing all scenarios assessed within the IPCC reporting process spanning the last decades. The analysis of scenario ensembles spanning four IPCC cycles provides a more dynamic overview of progress, allowing to track the evolution of baseline projections—as a proxy of climate effort effectiveness, improvements in modelling capacity and trends in low-carbon technoeconomic developments—and, in turn, of the solution space provided by IAMs to achieve predefined climate targets. Finally, we aim to understand how mitigation portfolios have evolved in IAMs and the driving factors behind this evolution. Results We have gathered, compiled and compared data from IAM scenarios included in four online scenario databases (Extended Data Table 1). These include AR5 published in 2014 24 , the Special Report on Global Warming of 1.5 °C (SR1.5) in 201825 and AR6 in 202226, which all benefit from a harmonized reporting template for all socioeconomic, energy system, emissions and climate outcomes, thereby allowing 25 30 35 40 45 50 55 60 65 SRES AR5 SR1.5 AR6 SRES AR5 SR1.5 AR6 SRES AR5 SR1.5 AR6 SRES AR5 SR1.5 AR6 2010202020302050 Gt fossil CO2 E&IP CO2 emissions in baseline scenarios and real-world observations GCP EDGAR a Lower interquartile range Median Higher interquartile range 34% 34% 33% 34% 287% 207% 205% 195% –185% –158% –163% –158% –32% 4% –8% –6% 0 50 100 150 200 250 300 350 SRES AR5 SR1.5 AR6 SRES AR5 SR1.5 AR6 SRES AR5 SR1.5 AR6 SRES AR5 SR1.5 AR6 Population GDP per capita Energy intensity of GDP CO2 intensity of energy E&IP CO2 emissions in 2050 relative to 2010 (%) Kaya decomposition, fossil CO2 in 2050 with respect to 2010 No-policy baselines b Fig. 1 | Analysis of E&IP CO2 emissions in baseline scenarios. a, Medians and interquartile ranges of emissions outcomes from no-policy baselines, for IPCC SRES, AR5, SR1.5 and AR6. Values for 2010 and 2020 include real-world observations by Global Carbon Project53 and EDGAR v.8.0 (ref. 54). For these observations, we took the average of 2009–2011 and of 2019 and 2021 for the value in 2010 and 2020, respectively, to avoid annual variability (including Covid-19 impacts) which most models do not account for. b, Kaya identity components for mean emission pathways, separately for SRES, AR5, SR1.5 and AR6. Values for SRES excluded industrial process emissions, and have been multiplied by 1.06 to be comparable with AR5, SR1.5 and AR6 scenarios, which commonly report the sum of E&IP emissions. Statistics are drawn from model means (Methods) and SRES n = 15, AR5 n = 53, SR1.5 n = 16, AR6 n = 47. Supplementary Table 1 gives an overview of included scenarios. Nature Climate Change | Volume 15 | February 2025 | 218–226 220 Analysis https://doi.org/10.1038/s41558-024-02198-6 whereas mitigation scenarios outline more ambitious transformations of sociotechnical and economic systems towards meeting ambitious climate goals. Baseline scenario projections for CO 2 emissions from energy and industrial processes (E&IP) in 2030 and 2050 have gradually decreased from SRES towards AR6 (Fig. 1a), reflecting transformation that has gradually happened (for example, due to policy efforts and low-carbon technology deployment in the historical years of models), evolution of modelling assumptions and/or capabilities (for example, relative costs and enhanced model representation of low-carbon technologies) or changes in assumed socioeconomic trajectories. Comparing scenario emissions in 2010 and 2020 with historical values shows that 2010 emissions were generally underestimated (below median) and 2020 emissions (2019–2021 average to largely avoid Covid-19 effects) overestimated in all datasets. Logically, older datasets show higher variability, as estimations are further away from model base years, allowing initial differences to increase over time. The improved baseline starting point, even in the absence of strong climate policies, and the clear overestimation of 2020 emissions, is compatible with recent literature pointing to futures of more than 4–5 °C end-of-century warming being increasingly unlikely15,29. Disentangling the 2010–2050 evolution of mean emission pathways of no-climate-policy baselines (Fig. 1b) in Kaya identity components (Methods) shows how differences in mean E&IP CO2 emissions (Fig. 1) are driven by a combination of causes. First, relative to AR5 and AR6, SRES projected significantly higher gross domestic product (GDP) growth, but also stronger decline in energy intensity of GDP and CO 2 intensity of energy, even in the absence of policy. The baseline emission reduction from AR5 to SR1.5 and AR6 is driven by a combination of all Kaya components, with again lower projected GDP growth playing a significant role, being the main driver behind AR5 scenarios overestimating emissions in 2020 30 . Overall, however, this overestimation of future GDP growth throughout the modelling ensembles may be driven by long-term uncertainty or structural assumptions in projections31 (Extended Data Fig. 1). Evolution of mitigation pathways When giving IAMs the objective to limit global temperature increase, radiative forcing or cumulative emissions to certain levels, they typically estimate the costor welfare-optimal path over time and space towards that level 32 . Nevertheless, model users determine from which year onwards IAMs must look for this path. AR5 scenarios assumed a level of mitigation activity by 2020 that did not eventuate, with the SR1.5–AR6 ensembles therefore deferring the onset of substantive mitigation activities until post-2020. The relatively delayed and higher peak in emissions resulted in deeper required emission cuts in 2050 in SR1.5–AR6 to be compatible with <1.5 °C warming (Fig. 2a). A similar but more moderate effect can be observed for 2 °C pathways, a pattern repeated for every region (Extended Data Fig. 2), confirming that it is a result of models attempting to find the optimal pathway to specified targets, rather than underlying region-specific developments. In terms of Kaya components (Methods), SR1.5 and AR6 include a much more substantial reduction in CO2 intensity of energy supply, than does AR5 (Fig. 2b), suggesting that these two ensembles rely heavily on supply-side technology innovation to compensate for the lagging mitigation response of the world. This observation may confirm criticisms of IPCC scenarios often relying on largely unprecedented technological change rather than slower economic growth or demand reduction for reaching climate objectives33, although there is also an incremental role for slower economic growth projections (AR6) or faster declining energy intensity of GDP (SR1.5) to make up for this delay relative to AR5. Interestingly, the role of technological carbon 0 5 10 15 20 25 30 35 40 2010 2020 2030 Year 2040 2050 MtCO2 from energy and industry 2 °C AR5 2 °C AR6 (including SR1.5) 1.5 °C AR5 1.5 °C AR6 (including SR1.5) Observed GCP Observed EDGAR 34% 33% 34% 207% 207% 196% –204% –211% –198% –41% –49% –54% –28% –27% –24% –100 –50 0 50 100 150 200 250 300 Kaya decomposition 2 °C compatible scenarios 34% 33% 34% 201% 205% 184% –214% –224% –199% –46% –69% –68% –33% –30% –30% –100 –50 0 50 100 150 200 250 300 AR5 SR1.5 AR6 AR5 SR1.5 AR6 AR5 SR1.5 AR6 AR5 SR1.5 AR6 AR5 SR1.5 AR6 Population GDP per capita Energy intensity of GDP CO2 intensity of energy CO2 capture and storage 1.5 °C compatible scenarios E&IP CO2 emissions in 2050 relative to 2010 (%) ab Fig. 2 | Analysis of E&IP CO2 emissions in mitigation pathways. a, Mean pathways in AR5 and AR6 from 2010 to 2050 and observed emissions in 2010 and 2020 (Fig. 1). Arrows in 2020 and 2050 reflect how the delayed mitigation in AR6 scenarios relative to AR5 scenarios leads to deeper required reductions in 2050 in AR6 scenarios to be compatible with the same climate target. Values for 2010 and 2020 include real-world observations by Global Carbon Project53 and EDGAR v.8.0 (ref. 54). b, Kaya identity components for 2 and 1.5 °C compatible scenarios, separately for AR5, SR1.5 and AR6. Statistics are drawn from model means (Methods) and Supplementary Table 1 gives an overview of included scenarios for each indicator. Nature Climate Change | Volume 15 | February 2025 | 218–226 221 Analysis https://doi.org/10.1038/s41558-024-02198-6 removals (such as CCS) slightly decreased from AR5 to AR6, despite higher mitigation pressure. The sectoral stocktake in mitigation efforts in 2010–2050 shows that the largest absolute contributions come from energy supply (covering all transformation stages—for example, mining, refining and electricity), but additional effort in moving from 2 to 1.5 °C pathways comes mostly from industry, buildings and transport (Fig. 3). Across scenario ensembles, there is a shift towards deeper mitigation in end-use sectors towards 2050, post-AR5, both for more modest 2 °C scenarios and more clearly for 1.5 °C scenarios, reflecting improved model incorporation of end-use technologies, such as electrification and hydrogen in industrial processes, battery-electric vehicles in transportation and electric heat pumps in buildings and/or the maturity of these technologies (for example, lower costs or growth constraints) post-201534,35. The evolution of CO 2 emissions from Energy Supply and AFOLU varies less in absolute quantity between scenario databases, but notably these sectors go towards net-negative emissions in SR1.5–AR6, instead of carbon neutrality (AR5). Energy-sector CO2 emissions become negative primarily as a result of electricity generation and fuel supply by BECCS, which first must offset residual sectoral emissions before becoming net-negative. Lower residual emissions in electricity generation (for example, more renewables) can facilitate lower emissions in the energy sector even if BECCS use is moderated. AFOLU CO2 emissions represent a balance between reforestation, afforestation, deforestation and—depending on model—forestry. There is a trade-off between BECCS and AFOLU because of limited land availability and saturation of AFOLU over time36. Energy technology mix and macroeconomic costs Focusing on mitigation scenarios compatible with 2 or 1.5 °C, results indicate gradual evolution of mitigation portfolios. In terms of primary energy mix (Fig. 4a), coal shares have dropped in more recent ensembles: while coal still played a non-negligible role in deep mitigation scenarios in AR5 (~10% of primary energy by 2050), its share is <10% in all 2 °C/1.5 °C scenarios in AR6. Reliance on natural gas was relatively stable among scenarios in AR5 (~20–25% of primary energy) but decreased in 1.5 °C scenarios in SR1.5–AR6 (towards ~15%). This drop is probably related to the need for deeper mitigation relative to AR5 (Fig. 2). Reliance on bioenergy for mitigation is also slightly lower in more recent scenarios, whereas reliance on solar and wind in the most ambitious scenarios increased from 10–15% (AR5) to 20–45% of total primary energy demand by 2050 (AR6), in response to rapidly declining technology costs and enhanced model representation of variable renewables (VRE) integration in the electricity system (for example, coupled with batteries). This last effect is clearly related to the electricity mix and significance in final energy demand, due to enhanced electrification (Fig. 4b). For example, the share of renewables in electricity increases from 40–50% (AR5) to 60–80% (SR1.5–AR6) by 2050 throughout scenarios, while nuclear drops from 15–20% to 10–15%, probably related with cost escalation and limited political support after the 2012 Fukushima event 37 . The share of electricity in final energy demand increases with deeper mitigation, but the overall level is ~10% higher in AR6 over AR5, leading to electrification rates of ~50% rather than ~40% for the most ambitious scenarios. Total primary and final energy use, both decreasing with increasing mitigation effort, have not significantly evolved since AR5 (Extended Data Fig. 3). AR5 2 °C AR5 1.5 °C SR1.5 2 °C SR1.5 1.5 °C AR6 2 °C AR6 1.5 °C –4 –2 0 2 4 6 8 10 12 14 16 18 2010 2030 Year YearYear Year Year 2050 Gt CO2 Energy supply 0 2 4 6 8 10 4 2010 2030 2050 Industry (combustion and processes) 0 1 2 3 2010 2030 2050 Buildings 0 2 4 6 8 10 2010 2030 2050 Transport Observed (EDGAR + GCP) –2 –1 0 1 2 3 4 5 2010 2030 2050 AFOLU a b –5 –10 0 5 10 15 20 25 30 35 40 45 AR5 SR1.5 AR6 AR5 SR1.5 AR6 AR5 SR1.5 AR6 AR5 SR1.5 AR6 2010* 2020* 2 °C compatible 1.5 °C compatible 2 °C compatible 1.5 °C compatible Observed emissions Means of 2030 projections Means of 2050 projections AFOLU Energy supply Buildings Transport Industry Fig. 3 | Global CO2 emissions by sector. Mean values for sectoral contribution of CO2 emissions in 2010 to 2050, separately for scenarios that are 2 and 1.5 °C compatible and for scenarios in IPCC AR5, SR1.5 and AR6. a,b, Absolute distribution per scenario (a) and evolution over time by sector (b). The asterisks indicate averages of observed values from EDGAR v.8.0 (ref. 54) (energy consumption sectors) and GCB53 (AFOLU sector) for 2009–2011 (as estimate for 2010) and of 2019 and 2021 (as estimate for 2020) to avoid annual variability (including Covid-19 impacts) which most models do not account for. Industry estimates from SR1.5 omitted because of inconsistent reporting of emissions from industrial processes. Statistics are drawn from model means (Methods) and Supplementary Table 1 gives an overview of included scenarios for each indicator. Nature Climate Change | Volume 15 | February 2025 | 218–226 222 Analysis https://doi.org/10.1038/s41558-024-02198-6 Generally, mitigation scenarios have evolved to depend less on CCS (Fig. 4c). The comparison of different CCS elements between scenario ensembles is not straightforward because of changes in the reporting structure of carbon sequestration, but the figure shows that reliance on CCS-powered fossil electricity significantly decreased from AR5 to SR1.5, staying relatively low but with high model variance in AR6, which is compatible with the abovementioned reduced reliance on fossil fuels and non-renewable electricity. Nevertheless, the overall role of CDR has not markedly reduced since AR5. While the role of BECCS marginally decreased, there has been a rise in new CDR technologies such as DACCS 38 (Extended Data Fig. 4). This distinction between CDR technologies is important, as it shows that, while mitigation has shifted towards renewables, the need for carbon removals has not declined with the need for deeper mitigation in SR1.5–AR6. Interestingly, SR1.5 scenarios appear to coincide more with AR6 trajectories for nearly all variables. The fact that 1.5 °C scenarios in SR1.5, as in AR6, required deeper mitigation compared to AR5 (Fig. 2), may play a role. However, for many variables (coal and nuclear shares, renewable electricity and electrification), the resemblance clearly starts with 2 °C scenarios already, and the similar tendency of 2 °C scenarios throughout the IPCC scenario ensembles (Fig. 2) suggests that there are more structural causalities too; for example, updated model structures and/or assumptions. For decades now, the IPCC has been tasked with assessing the literature on macroeconomic implications of reducing emissions and has responded by publishing both estimates of, and limitations inherent in, long-term macroeconomic projections 39 . Acknowledging the numerous methodological issues in these estimates, as well as inherent differences in how models address the economics of mitigation, we track how assessments of macroeconomic consequences have evolved from AR5 to AR6 (Fig. 4d). We observe that, for 2 °C scenarios, estimates have shifted from reductions of 2–3% of economic activity (GDP) in AR5 to 1–2% in AR6. For 1.5 °C scenarios, these differences are smaller, dropping from 3–5% to 2–4% of total GDP. Nevertheless, considering the deeper mitigation requirements in 2050 in AR6 relative to AR5 (Fig. 2), the observation that macroeconomic costs are lower in AR6 signals that a broader suite of mitigation options was 0 1 2 3 4 5 6 7 8 9 1,500–1,600 1,400–1,500 1,300–1,400 1,200–1,300 1,100–1,200 1,000–1,100 900–1,000 800–900 700–800 Percentage loss in output due to climate policy Cumulative GtCO2 2010–2050 Macro-economic output loss by 2050 0 10 20 30 40 Percentage of total primary energy Coal share Bio-energy share Gas share Wind and solar share 0 10 20 30 40 50 60 Share of renewables in electricity mix (%) Electrification of final energy use 0 10 20 30 40 50 60 70 80 90 100 Share of renewables in electricity mix (%) Renewable electricity 0 5 10 15 20 25 30 35 Share of renewables in electricity mix (%) Nuclear electricity 0 2 4 6 8 10 12 14 16 18 20 1,500–1,600 1,400–1,500 1,300–1,400 1,200–1,300 1,100–1,200 1,000–1,100 900–1,000 800–900 700–800 GtCO2 captured Cumulative GtCO2 2010–2050 Carbon capture and storage (including DACCS) Primary energy shares in 2050 Electricity indicators in 2050 2 °C compatible 1.5 °C compatible 0 10 20 30 40 50 60 70 1,500–1,600 1,400–1,500 1,300–1,400 1,200–1,300 1,100–1,200 1,000–1,100 900–1,000 800–900 700–800 EJ electricity generation Cumulative GtCO2 2010–2050 Fossil electricity with CCS 1.5 °C compatible 2 °C compatible 0 5 10 15 20 25 30 1,500–1,600 1,400–1,500 1,300–1,400 1,200–1,300 1,100–1,200 1,000–1,100 900–1,000 800–900 700–800 EJ electricity generation Cumulative GtCO2 2010–2050 Bio-electricity with CCS 2 °C compatible 1.5 °C compatible Carbon capture indicators in 2050 2 °C compatible 1.5 °C compatible a b c d AR5 SR1.5 AR6 Fig. 4 | Energy system and mitigation indicators by 2050. a–d, Interquartile ranges of structural indicators (primary energy shares (a), electricity indicators (b), carbon capture indicators (c) and macroeconomic output loss in (d)) of mitigation pathways by 2050, plotted separately for IPCC AR5, SR1.5 and AR6 as a function of mitigation effort in terms of cumulative fossil CO2 emissions from 2010 to 2050, marked by 2 °C and 1.5 °C compatibility. The metric for macroeconomic output loss (d) differs between reported scenarios, and includes GDP losses, consumption losses or the area under the MAC curve. Statistics are drawn from model means (Methods) and Supplementary Table 1 gives an overview of included scenarios. EJ, exajoule. Nature Climate Change | Volume 15 | February 2025 | 218–226 223 Analysis https://doi.org/10.1038/s41558-024-02198-6 available in AR6, assumptions over future mitigation technology costs have become more optimistic and/or structural changes in IAMs have lowered the resistance to major shifts in energy systems. Interestingly, the macroeconomic implications in SR1.5 largely resembled those of AR6 for 2 °C scenarios, but those of AR5 for 1.5 °C scenarios. This may signal that assumptions about major energy transformations may have been more optimistic in SR1.5 relative to AR5, but model capabilities for reaching deep mitigation were not yet as well-developed as in AR6. Note that these costs do not include climate impacts, and the economic impacts of climate inaction as outlined in IPCC WGII may well surpass the IAM-reported costs of mitigation39. The figures in this study show a gradually evolving emission/temperature future (in no-climate-policy baselines) and transformation path for reaching 2 °C and 1.5 °C futures, as well as increasing mitigation stringency required for nearly all sectors and regions to achieve 1.5 oC. We discuss likely technological drivers (Box 1) and political funding-related drivers (Box 2) for these changes. Discussion IAM research has seen an impressive uplift during the past two decades. Combined with consistent scenario reporting 40 , this allows for a multidimensional tracking of how transformation pathways have evolved. Our analysis takes aim at this evolution over the last decade through a detailed analysis of thousands of IAM scenarios in IPCC scenario databases. There is continuous evolution of most analysed indicators over scenario ensembles with a clear direction towards increasing BOX 1 The role of technology in evolving mitigation scenarios Several clean energy technologies have been scaled up significantly over the past decade, with cost reduction and performance improvements unforeseen at this scale in modelled scenarios, which have been gradually implemented in more recent scenarios. The clearest example is solar PV, for which the cost assumption for 2020 was three to six times lower in AR6 than in AR5 (interquartile ranges), and observed costs around 2020 according to IRENA (2021)55 were lower than assumed costs for 2050 in AR5 and SR1.5 scenarios. Similarly, for technologies such as onshore and offshore wind, battery storage, electric vehicles, hydrogen electrolysers and electric heat pumps, recent cost reductions and/ or performance improvements comfortably exceeded modelled assumptions in earlier model runs and IPCC reports. The updated representation of cheaper-than-expected low-carbon technologies in models causes projected baseline emissions to be lower, as clean technologies compete against dirty alternatives even in the absence of climate policy (Fig. 1), and mitigation technologies to shift radically with the same mitigation target (Fig. 4). 7,000 6,000 5,000 AR5 AR6 AR6 SR1.5 SR1.5 AR5 Capital costs—solar PV US$(2010)/kWe IRENA (2021) observed costs 2010, 2015, 2020 Capital costs—wind onshore 4,000 3,000 2,000 2010 2030 Year Year 2050 2010 2030 2050 1,000 0 Moreover, the representation of end-use sector technologies has been substantially improved from AR5 towards AR6. This continuous evolution of IAMs also opens new opportunities to mitigation43,47, and can be expected to continue into future scenario exercises. Recent industry-specific modelling exercises with higher sectoral detail achieve larger emission reductions than IAM scenarios considered in the IPCC cycles assessed in this article56,57, to the point that some sectors previously considered as hard to abate may now have the option of a much steeper mitigation trajectory, such as the iron and steel sector with new steel-related investment in the European Union (EU) already dominated by low-carbon options (for example, electric arc furnace and hydrogen-based direct reduction of iron). We expect that these recent developments will also be reflected in the new generation of IAM scenarios. BOX 2 Scenarios from European modelling teams dominate IPCC assessed scenario database, shaping energy transition patterns The changing patterns we found in the indicators examined in this study (Fig. 4) were especially accentuated in scenarios developed by organizations from Europe: AR6 scenarios from European organizations had a higher share of solar PV and wind power in electricity generation and primary energy as well as higher electrification than the rest—and less reliance on nuclear and CCS (Supplementary Fig. 1). These differences were less pronounced in AR5 and SR1.5 and were even reversed for bioenergy. Similar patterns were found by comparing scenarios of different warming levels between European and other organizations (Supplementary Figs. 2–4), with 2 °C scenarios showing the largest differences. While scenario developers from Europe were also using lower cost assumptions in AR6 for solar and wind than the rest (Supplementary Fig. 5), technological costs did not seem to be the only reason for these differences. For instance, even when controlling for the cost assumptions of solar PV and wind power, the share of renewables in primary energy (Supplementary Fig. 1) was still statistically significantly higher in scenarios coming from European organizations than the rest (median of 25% versus 10%, one-way analysis of covariance, F = 56.49, P < 0.001). Similarly, ref. 3 showed that projected solar PV growth was higher in AR5 and SR1.5 scenarios from European organizations compared to scenarios from Asian or North American organizations. European scenario developers have contributed most scenarios to the examined reports (71% in AR6, 75% in SR1.5 and 66% in AR5; Supplementary Table 3), so they expectedly have a large impact on the overall patterns in the IPCC scenario databases. Since their share in the databases has not changed markedly from AR5, the significant evolution observed here may be due to an increased use of scenario designs that explore higher shares of renewables and/or feasibility concerns58. Nature Climate Change | Volume 15 | February 2025 | 218–226 224 Analysis https://doi.org/10.1038/s41558-024-02198-6 electrification rates and higher VRE shares in electricity generation. Simultaneously, the use of nuclear, bioenergy and CCS with fossil energy has declined in more recent scenarios, despite more stringent 2050 targets. Since 2010, remaining carbon budgets have shrunk by ~400 GtCO 2 (~0.5 °C), yet the achievability and macroeconomic implications of delivering the same climate target have not markedly changed. An important interpretation of this is that AR5 portrayed more ambitious climate action as harder and costlier than nowadays calculated. The SR1.5–AR6 scenario ensembles feature more ambitious emissions cuts in energy end-use sectors (industry, buildings and transport), while allowing energy supply and AFOLU to reach net-negative CO2 emissions. The main causes we found behind the explored differences between scenario ensembles also imply lessons for future scenario development. A major driver lies in faster-than-expected reduction in clean technology costs and improved representation of critical end-use technologies (Box 1). The deviation between projected and observed trends in many key technologies41,42 points to the need for sustained effort to further update models. Any failure by IPCC scenarios to adequately reflect the practicability/plausibility of certain changes becomes more significant with declining carbon budgets. It is, therefore, important that modelling communities work closely with stakeholders to keep their critical technology assumptions up-to-date amidst rapidly evolving market dynamics. Since CDR levels in mitigation scenarios are dependent on the emission reduction levers in hard-to-abate sectors included in models 43 , such developments may reduce reliance on CDR. A second cause we identified was political (Box 2): narratives stemming from a European context may be more pro-renewables given the rapid roll-out of solar and wind envisioned in the European Green Deal44 and, perhaps, the EU being a major funder of such studies (for example, via H2020 and Horizon Europe; Supplementary Table 3)18. The dominance of scenarios from European institutes in scenario ensembles may explain the increased focus on VRE in more recent scenario ensembles. European narratives may, however, contrast with narratives from developing economies, such as India, where mitigation strategies are interwoven with socioeconomic development45. High economic growth projections have historically been used as an argument to delay mitigation, as future generations would be more capable of paying for adaptation and/or CDR technologies to regulate temperatures46. Although most scenarios in the explored ensembles are not based on intergenerational cost–benefit analysis, the observation that projections for future economic growth have been continuously corrected downwards (Fig. 1 and Extended Data Fig. 1) may imply reduced innovation potential for breakthroughs and value added of CDR technologies31. Such technologies, therefore, increase future costs of both mitigation (through later action) and adaptation (through temperature overshoot) 47 , while also involving other risks to future societies13. The track record of real-world technology costs outperforming forecasts (Box 1) illustrates that lacking forecasts of inputs to mitigation scenarios have falsely favoured delayed action throughout the past decade. However tempting to conclude that renewables in future low-carbon electricity projections will continue to increase, owing to spectacular cost reductions and deployment relative to CCS and nuclear, this is not certain. Grid balancing, social acceptance, mineral supply chains and so on must be considered as potential offsets to continued capital-cost declines. The future direction of other observed trends is also unclear; notably, although projected mitigation costs have fallen (or, for 1.5 °C pathways, largely remained unchanged) with decreasing technology costs more-than-offsetting shrinking carbon budgets, this balance can only last so long. Inevitably, especially for 1.5 °C pathways, mitigation costs will rise, as the remaining carbon budget eventually diminishes and emissions removals become central to keeping below this temperature threshold. A clearer trend, from AR5 to AR6, is the reduction in coal and gas shares in future primary energy demand in the lowest carbon budget scenarios. This could well continue into AR7 but, regardless, near-term geopolitical developments must be factored in—for example, potential lock—into LNG infrastructure in light of a wave of new global investment48. It is likely that the 1.5 °C temperature stabilization goal will be exceeded within the next 5–10 years49, dictating a reframing of scenarios and IAMs. Instead of more elaborate ways to keep below 1.5 °C, focus may shift towards feasibility50, requiring more critical thinking on how different technologies may be deployed under different policy and regional contexts. This emphasizes the importance of using multiple streams of evidence when using/interpreting scenarios, such as sectoral and national expertise. Models and scenarios are enticing because of their ability to integrate hundreds or thousands of assumptions, but this also makes them susceptible to providing a false sense of confidence in results. The surge of the ‘net-zero’ concept, stemming from physical climate science51, and increasingly applied in political pledges and scenario design50, had no role in AR5, a relatively small role in AR6, but will probably be instrumental towards AR7, by itself conditioning several indicators explored in this study. The world now has a multitude of mitigation policies in place and, despite lack of a strong tradition of such analysis, their impact will complicate the development of realistic ‘reference scenarios’ in future exercises. First, without real-world comparison, evaluating no-climate-policy baselines is not as meaningful at the global level15. Second, despite a growing number of relevant studies 14,50,52 , pathways reflecting current policies and/or Nationally Determined Contributions are also hard to evaluate, since policy constantly changes. Third, backcasting scenarios typically performed to determine how the world could transform and achieve ambitious climate targets mostly apply idealized assumptions, such as globally harmonized carbon prices across sectors/regions, a situation increasingly unlikely to be realized. Evaluating scenarios will, thus, require an increasing degree of expert judgement to elicit policy relevant recommendations. Online content Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41558-024-02198-6. References 1. IPCC Climate Change 2023: Synthesis Report (eds Core Writing Team, Lee, H. & J. Romero, J.) (IPCC, 2023). 2. IPCC Climate Change 2014: Synthesis Report (eds Core Writing Team, Pachauri, R. K. & Meyer, L. A.) (IPCC, 2014). 3. Jaxa-Rozen, M. & Trutnevyte, E. Sources of uncertainty in long-term global scenarios of solar photovoltaic technology. Nat. Clim. Change 11, 266–273 (2021). 4. Binsted, M. An electrified road to climate goals. Nat. Energy 7, 9–10 (2022). 5. Luderer, G. et al. Impact of declining renewable energy costs on electrification in low-emission scenarios. Nat. Energy 7, 32–42 (2021). 6. Strauch, Y. Beyond the low-carbon niche: global tipping points in the rise of wind, solar, and electric vehicles to regime scale systems. Energy Res. Soc. Sci. 62, 101364 (2020). 7. Bauer, C. et al. Charging sustainable batteries. Nat. Sustain. 5, 176–178 (2022). 8. Perdana, S. et al. Expert perceptions of game-changing innovations towards net zero. Energy Strateg. Rev. 45, 101022 (2023). Nature Climate Change | Volume 15 | February 2025 | 218–226 225 Analysis https://doi.org/10.1038/s41558-024-02198-6 9. Lane, J., Greig, C. & Garnett, A. Uncertain storage prospects create a conundrum for carbon capture and storage ambitions. Nat. Clim. Change 11, 925–936 (2021). 10. Anderson, K. & Peters, G. The trouble with negative emissions. Science 354, 182–183 (2016). 11. Fuss, S. et al. Negative emissions—Part 2: Costs, potentials and side effects. Environ. Res. Lett. 13, 063002 (2018). 12. Fridahl, M. & Lehtveer, M. Bioenergy with carbon capture and storage (BECCS): global potential, investment preferences, and deployment barriers. Energy Res. Soc. Sci. 42, 155–165 (2018). 13. Smith, W. Pandora’s Toolbox: The Hopes and Hazards of Climate Intervention (Cambridge Univ. Press, 2022). 14. Sognnaes, I. et al. A multi-model analysis of long-term emissions and warming implications of current mitigation efforts. Nat. Clim. Change 11, 1055–1062 (2021). 15. Hausfather, Z. & Peters, G. P. Emissions—the ‘business as usual’ story is misleading. Nature 577, 618–620 (2020). 16. Boon-Falleur, M., Grandin, A., Baumard, N. & Chevallier, C. Leveraging social cognition to promote effective climate change mitigation. Nat. Clim. Change 12, 332–338 (2022). 17. Koasidis, K., Nikas, A. & Doukas, H. Why integrated assessment models alone are insufficient to navigate us through the polycrisis. One Earth 6, 205–209 (2023). 18. Nikas, A. et al. Perspective of comprehensive and comprehensible multi-model energy and climate science in Europe. Energy 215, 119153 (2021). 19. Weber, C. et al. Mitigation scenarios must cater to new users. Nat. Clim. Change 8, 845–848 (2018). 20. van Beek, L., Hajer, M., Pelzer, P., van Vuuren, D. & Cassen, C. Anticipating futures through models: the rise of Integrated Assessment Modelling in the climate science-policy interface since 1970. Glob. Environ. Change 65, 102191 (2020). 21. Braunreiter, L., Marchand, C. & Blumer, Y. Exploring possible futures or reinforcing the status-quo? The use of model-based scenarios in the Swiss energy industry. Renew. Sustain. Energy Transit. 3, 100046 (2023). 22. Guivarch, C. et al. Using large ensembles of climate change mitigation scenarios for robust insights. Nat. Clim. Change 12, 428–435 (2022). 23. Dekker, M. M. et al. Spread in climate policy scenarios unravelled. Nature 624, 309–316 (2023). 24. IPCC Climate Change 2014: Mitigation of Climate Change (eds Edenhofer, O. et al.) (Cambridge Univ. Press, 2014). 25. Huppmann, D. et al. IAMC 1.5 °C Scenario Explorer and Data Hosted by IIASA (IIASA, 2018); https://doi.org/10.22022/SR15/ 08-2018.15429 26. Byers, E. et al. AR6 scenarios database (1.0) [Data set]. Zenodo https://doi.org/10.5281/zenodo.5886912 (2022). 27. IPCC Special Report on Emissions Scenarios (eds Nakicenovic, E. & Swart, R.) (Cambridge Univ. Press, 2000). 28. Hermwille, L., Obergassel, W. & Arens, C. The transformative potential of emissions trading. Carbon Manag. 6, 261–272 (2015). 29. Pielke, R., Burgess, M. G. & Ritchie, J. Plausible 2005–2050 emissions scenarios project between 2 °C and 3 °C of warming by 2100. Environ. Res. Lett. 17, 024027 (2022). 30. Burgess, M. G., Ritchie, J., Shapland, J. & Pielke, R. IPCC baseline scenarios have over-projected CO2 emissions and economic growth. Environ. Res. Lett. 16, 014016 (2021). 31. Burgess, M. G. et al. Multidecadal dynamics project slow 21st-century economic growth and income convergence. Commun. Earth Environ. 4, 220 (2023). 32. Gambhir, A., Ganguly, G. & Mittal, S. Climate change mitigation scenario databases should incorporate more non-IAM pathways. Joule 6, 2663–2667 (2022). 33. Keyßer, L. T. & Lenzen, M. 1.5 °C degrowth scenarios suggest the need for new mitigation pathways. Nat. Commun. 12, 2676 (2021). 34. Grubler, A. et al. A low energy demand scenario for meeting the 1.5 °C target and sustainable development goals without negative emission technologies. Nat. Energy 3, 515–527 (2018). 35. Keppo, I. et al. Exploring the possibility space: taking stock of the diverse capabilities and gaps in integrated assessment models. Environ. Res. Lett. 16, 053006 (2021). 36. Humpenöder, F. et al. Investigating afforestation and bioenergy CCS as climate change mitigation strategies. Environ. Res. Lett. 9, 064029 (2014). 37. Nuclear Power and Secure Energy Transitions (IEA, 2022); www.iea.org/reports/nuclear-power-and-secure-energytransitions 38. Realmonte, G. et al. An inter-model assessment of the role of direct air capture in deep mitigation pathways. Nat. Commun. 10, 3277 (2019). 39. Köberle, A. C. et al. The cost of mitigation revisited. Nat. Clim. Change 11, 1035–1045 (2021). 40. Cointe, B. The AR6 scenario explorer and the history of IPCC scenarios databases: evolutions and challenges for transparency, pluralism and policy-relevance. npj Clim. Action 3, 3 (2024). 41. He, G. et al. Rapid cost decrease of renewables and storage accelerates the decarbonization of China’s power system. Nat. Commun. 11, 2486 (2020). 42. Way, R., Ives, M. C., Mealy, P. & Farmer, J. D. Empirically grounded technology forecasts and the energy transition. Joule 6, 2057–2082 (2022). 43. Edelenbosch, O. Y. et al. Reducing sectoral hard-to-abate emissions to limit reliance on carbon dioxide removal. Nat. Clim. Change 14, 715–722 (2024). 44. 2030 Climate Target Plan (European Commission, 2021); https://ec.europa.eu/clima/euaction/european-green-deal/ 2030-climate-target-plan_en#documents 45. Sudharmma Vishwanathan, S., Fragkos, P., Fragkiadakis, K. & Garg, A. Assessing enhanced NDC and climate compatible development pathways for India. Energy Strateg. Rev. 49, 101152 (2023). 46. Nordhaus, W. The Climate Casino: Risk, Uncertainty, and Economics for a Warming World (Yale Univ. Press, 2013). 47. Fuhrman, J. et al. Ambitious efforts on residual emissions can reduce CO2 removal and lower peak temperatures in a net-zero future. Environ. Res. Lett. 19, 064012 (2024). 48. World Energy Investment 2024 (IEA, 2024); www.iea.org/reports/ world-energy-investment-2024 49. Forster, P. M. et al. Indicators of global climate change 2023: annual update of key indicators of the state of the climate system and human influence. Earth Syst. Sci. Data 16, 2625–2658 (2024). 50. van de Ven, D.-J. et al. A multimodel analysis of post-Glasgow climate targets and feasibility challenges. Nat. Clim. Change 13, 570–578 (2023). 51. Allen, M. R. et al. Net zero: science, origins, and implications. Annu. Rev. Environ. Resour. 47, 849–887(2022). 52. Roelfsema, M. et al. Taking stock of national climate policies to evaluate implementation of the Paris Agreement. Nat. Commun. 11, 2096 (2020). 53. Friedlingstein, P. et al. Global carbon budget 2023. Earth Syst. Sci. Data 15, 5301–5369 (2023). 54. Crippa, M. et al. GHG Emissions of All World Countries (Publications Office of the European Union, 2023). 55. Renewable Power Generation Costs in 2020 (IRENA, 2021). Nature Climate Change | Volume 15 | February 2025 | 218–226 226 Analysis https://doi.org/10.1038/s41558-024-02198-6 56. Bataille, C. et al. Towards net-zero emissions concrete and steel in India, Brazil and South Africa. Clim. Policy https://doi.org/10.1080/ 14693062.2023.2187750 (2023). 57. Low-Carbon Technologies for the Global Steel Transformation. A Guide to the Most Effective Ways to Cut Emissions in Steelmaking (Agora Industry, 2024); www.agora-industry.org/fileadmin/ Projekte/2021/2021-06_IND_INT_GlobalSteel/A-IND_324_ Low-Carbon-Technologies_WEB.pdf 58. Riahi, K. et al. Cost and attainability of meeting stringent climate targets without overshoot. Nat. Clim. Change 11, 1063–1069 (2021). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. © The Author(s), under exclusive licence to Springer Nature Limited 2025 Nature Climate Change Analysis https://doi.org/10.1038/s41558-024-02198-6 Extended Data Table 1 | Details of scenario databases included in the comparative analysis IPCC report name Temporal coverage of Global scenario count Global model count Source SRES Special Report on Emissions Scenarios ~2000 80 10 (IPCC, 2000) AR5 5th Assessment report 2010-2014 1184 31 (IPCC, 2014) SR1.5 Special Report 1.5°C 2015-2018 416 25 (IPCC, 2018) AR6 6th Assessment report 2015-2021 (includes those in SR1.5) 2304 95 (Byers et al., 2022)