Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

Using deep learning to generate key variables in global mitigation scenarios

Nature Climate Change (2025)Cite this article

Subjects

Abstract

Integrated assessment models (IAMs) are the dominant tools for projecting mitigation scenarios. However, IAM-based scenarios often face challenges such as modelling biases and large computational burden. Here we develop a deep learning framework to generate key variables through synthetic mitigation scenarios aligned with the Sixth Assessment Report (AR6) Scenarios Database. By analysing 1,202 scenarios from a diverse set of IAMs, we select key drivers that enable a more detailed sectoral representation. Next, we trained three generative deep learning models to produce 30,000 synthetic scenarios at low computational cost across various IPCC AR6 climate categories, replicating variable distributions and correlations while also demonstrating physical consistency in power sector variables through internal validation checks. We found that the variational autoencoder achieved the highest label transferring accuracy among three frameworks. This study illustrates the potential of deep learning to complement IAM approaches and provides a basis for handling complex mitigation scenario generation tasks.

Access through your institution
Buy or subscribe

This is a preview of subscription content, access via your institution

Access options

Access through your institution

Buy this article

  • Purchase on SpringerLink
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Overview of research framework.
Fig. 2: Identification of key determinants of mitigation scenarios.
Fig. 3: IPCC-consistent synthetic scenarios generated by VAE.
Fig. 4: Correlations among key variables in AR6 scenarios and VAE-generated synthetic scenarios for the C1234 category.
Fig. 5: Internal consistency and computational performance of synthetic scenarios.

Similar content being viewed by others

Data availability

The AR6 Scenarios Database (ref. 27) is publicly available at https://iiasa.ac.at/models-tools-data/ar6-scenario-explorer-and-database. VAE-generated scenarios (n = 30,000) are available via Zenodo at https://doi.org/10.5281/zenodo.15240553 (ref. 74). Source data are provided with this paper.

Code availability

Source code to reproduce this analysis is available via Zenodo at https://doi.org/10.5281/zenodo.15280653 (ref. 75).

References

  1. IPCC: Summary for Policymakers. In Climate Change 2023: Synthesis Report (eds Core Writing Team et al.) (IPCC, 2023).

  2. Hoegh-Guldberg, O. et al. The human imperative of stabilizing global climate change at 1.5 °C. Science 365, eaaw6974 (2019).

    Article  CAS  Google Scholar 

  3. Lade, S. J. et al. Human impacts on planetary boundaries amplified by Earth system interactions. Nat. Sustain. 3, 119–128 (2020).

    Article  Google Scholar 

  4. Riahi, K. et al. The Shared Socioeconomic Pathways and their energy, land use, and greenhouse gas emissions implications: an overview. Glob. Environ. Change 42, 153–168 (2017).

    Article  Google Scholar 

  5. Luderer, G. et al. Impact of declining renewable energy costs on electrification in low-emission scenarios. Nat. Energy 7, 32–42 (2022).

    Article  Google Scholar 

  6. DeAngelo, J. et al. Energy systems in scenarios at net-zero CO2 emissions. Nat. Commun. 12, 6096 (2021).

    Article  CAS  Google Scholar 

  7. McCollum, D. L. et al. Energy investment needs for fulfilling the Paris Agreement and achieving the Sustainable Development Goals. Nat. Energy 3, 589–599 (2018).

    Article  Google Scholar 

  8. Arneth, A., Brown, C. & Rounsevell, M. Global models of human decision-making for land-based mitigation and adaptation assessment. Nat. Clim. Change 4, 550–557 (2014).

    Article  Google Scholar 

  9. Chowdhury, A. K. et al. Hydropower expansion in eco-sensitive river basins under global energy-economic change. Nat. Sustain. 7, 213–222 (2024).

    Article  Google Scholar 

  10. Ou, Y. et al. Can updated climate pledges limit warming well below 2 °C? Science 374, 693–695 (2021).

    Article  CAS  Google Scholar 

  11. Clarke, L. et al. in IPCC Climate Change 2014: Mitigation of Climate Change (eds Edenhofer, O. et al.) Ch. 6 (Cambridge Univ. Press, 2014).

  12. Rogelj, J. et al. A new scenario logic for the Paris Agreement long-term temperature goal. Nature 573, 357–363 (2019).

    Article  CAS  Google Scholar 

  13. Weyant, J. Some contributions of integrated assessment models of global climate change. Rev. Environ. Econ. Policy 11, 115–137 (2017).

  14. Weyant, J. P. A perspective on integrated assessment. Clim. Change 95, 317–323 (2009).

    Article  Google Scholar 

  15. Krey, V. Global energy‐climate scenarios and models: a review. Wiley Interdiscip. Rev.: Energy Environ. 3, 363–383 (2014).

    Google Scholar 

  16. Wang, H. et al. Scientific discovery in the age of artificial intelligence. Nature 620, 47–60 (2023).

    Article  CAS  Google Scholar 

  17. Kanyako, F. et al. Compounding uncertainties in economic and population growth increase tail risks for relevant outcomes across sectors. Earth’s Future 12, e2023EF003930 (2024).

    Article  Google Scholar 

  18. Gillingham, K. et al. Modeling uncertainty in integrated assessment of climate change: a multimodel comparison. J. Assoc. Environ. Reso. 5, 791–826 (2018).

    Google Scholar 

  19. Riahi, K. et al. Cost and attainability of meeting stringent climate targets without overshoot. Nat. Clim. Change 11, 1063–1069 (2021).

    Article  Google Scholar 

  20. O’Neill, B. C. et al. Achievements and needs for the climate change scenario framework. Nat. Clim. Change 10, 1074–1084 (2020).

    Article  Google Scholar 

  21. O’Neill, B. C. et al. The scenario model intercomparison project (ScenarioMIP) for CMIP6. Geosci. Model Dev. 9, 3461–3482 (2016).

    Article  Google Scholar 

  22. Rosenzweig, C. et al. The agricultural model intercomparison and improvement project (AgMIP): protocols and pilot studies. Agric. For. Meteorol. 170, 166–182 (2013).

    Article  Google Scholar 

  23. 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).

    Article  Google Scholar 

  24. Harmsen, M. et al. Integrated assessment model diagnostics: key indicators and model evolution. Environ. Res. Lett. 16, 054046 (2021).

    Article  Google Scholar 

  25. Guivarch, C. et al. Using large ensembles of climate change mitigation scenarios for robust insights. Nat. Clim. Change 12, 428–435 (2022).

    Article  Google Scholar 

  26. Huppmann, D. et al. IAMC 1.5 °C scenario explorer and data hosted by IIASA. Integrated Assessment Modeling Consortium & International Institute for Applied Systems Analysis https://doi.org/10.22022/SR15/08-2018.15429 (2018).

  27. Byers, E. et al. AR6 Scenarios Database. Zenodo https://doi.org/10.5281/zenodo.7197970 (2022).

  28. Dekker, M. M. et al. Spread in climate policy scenarios unravelled. Nature 624, 309–316 (2023).

    Article  CAS  Google Scholar 

  29. IPCC: Summary for Policymakers. In Climate Change 2022: Mitigation of Climate Change (eds Shukla, P. R. et al.) (Cambridge Univ. Press, 2022).

  30. Guivarch, C. et al. Annex III: Scenarios and Modelling Methods (IPCC, 2022).

  31. 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).

    Article  Google Scholar 

  32. Dekker, M. M. et al. Identifying energy model fingerprints in mitigation scenarios. Nat. Energy 8, 1395–1404 (2023).

    Article  Google Scholar 

  33. van Vuuren, D. P. et al. The representative concentration pathways: an overview. Clim. Change 109, 5–31 (2011).

    Article  Google Scholar 

  34. Keyßer, L. T. & Lenzen, M. 1.5 °C degrowth scenarios suggest the need for new mitigation pathways. Nat. Commun. 12, 2676 (2021).

    Article  Google Scholar 

  35. Gambhir, A., Ganguly, G. & Mittal, S. Climate change mitigation scenario databases should incorporate more non-IAM pathways. Joule 6, 2663–2667 (2022).

    Article  Google Scholar 

  36. Hickel, J. et al. Urgent need for post-growth climate mitigation scenarios. Nat. Energy 6, 766–768 (2021).

    Article  Google Scholar 

  37. Pirani, A. et al. Scenarios in IPCC assessments: lessons from AR6 and opportunities for AR7. npj Clim. Action 3, 1 (2024).

    Article  Google Scholar 

  38. Sengupta, S. et al. A review of deep learning with special emphasis on architectures, applications and recent trends. Knowl.-Based Syst. 194, 105596 (2020).

    Article  Google Scholar 

  39. Karniadakis, G. E. et al. Physics-informed machine learning. Nat. Rev. Phys. 3, 422–440 (2021).

    Article  Google Scholar 

  40. Qiu, J., Wu, Q., Ding, G., Xu, Y. & Feng, S. A survey of machine learning for big data processing. EURASIP J. Adv. Signal Process. 2016, 67 (2016).

    Article  Google Scholar 

  41. Buster, G., Benton, B. N., Glaws, A. & King, R. N. High-resolution meteorology with climate change impacts from global climate model data using generative machine learning. Nat. Energy 9, 894–906 (2024).

  42. Price, I. et al. Probabilistic weather forecasting with machine learning. Nature 637, 84–90 (2025).

    Article  CAS  Google Scholar 

  43. Grossmann, I. et al. AI and the transformation of social science research. Science 380, 1108–1109 (2023).

    Article  CAS  Google Scholar 

  44. Kingma, D. P. & Welling, M. Auto-encoding variational bayes. In 2nd International Conference on Learning Representations (2014).

  45. Goodfellow, I. et al. Generative adversarial nets. Adv. Neural Inf. Process. Syst. 27, 2672–2680 (2014).

    Google Scholar 

  46. Mirza, M. & Osindero, S. Conditional generative adversarial nets. Preprint at https://arxiv.org/abs/1411.1784 (2014).

  47. Li, Q. et al. Coupled GAN with relativistic discriminators for infrared and visible images fusion. IEEE Sens. J. 21, 7458–7467 (2019).

    Article  Google Scholar 

  48. Rasp, S., Pritchard, M. S. & Gentine, P. Deep learning to represent subgrid processes in climate models. Proc. Natl Acad. Sci. USA 115, 9684–9689 (2018).

    Article  CAS  Google Scholar 

  49. Scher, S. Toward data‐driven weather and climate forecasting: approximating a simple general circulation model with deep learning. Geophys. Res. Lett. 45, 12,616–12,622 (2018).

    Article  Google Scholar 

  50. Bi, K. et al. Accurate medium-range global weather forecasting with 3D neural networks. Nature 619, 533–538 (2023).

    Article  CAS  Google Scholar 

  51. Lam, R. et al. Learning skillful medium-range global weather forecasting. Science 382, 1416–1421 (2023).

    Article  CAS  Google Scholar 

  52. Khodayar, M., Liu, G., Wang, J. & Khodayar, M. E. Deep learning in power systems research: a review. CSEE J. Power Energy Syst. 7, 209–220 (2020).

    Google Scholar 

  53. Ozcanli, A. K., Yaprakdal, F. & Baysal, M. Deep learning methods and applications for electrical power systems: a comprehensive review. Int. J. Energy Res. 44, 7136–7157 (2020).

    Article  Google Scholar 

  54. Wang, H., Lei, Z., Zhang, X., Zhou, B. & Peng, J. A review of deep learning for renewable energy forecasting. Energy Convers. Manag. 198, 111799 (2019).

    Article  Google Scholar 

  55. Aslam, S. et al. A survey on deep learning methods for power load and renewable energy forecasting in smart microgrids. Renew. Sustain. Energy Rev. 144, 110992 (2021).

    Article  Google Scholar 

  56. Almalaq, A. & Edwards, G. A review of deep learning methods applied on load forecasting. In 16th IEEE International Conference on Machine Learning and Applications (ICMLA) 511–516 (IEEE, 2017).

  57. Callaghan, M. et al. Machine-learning-based evidence and attribution mapping of 100,000 climate impact studies. Nat. Clim. Change 11, 966–972 (2021).

    Article  Google Scholar 

  58. Rolnick, D. et al. Tackling climate change with machine learning. ACM Comput. Surv. 55, 42 (2022).

    Google Scholar 

  59. Clutton-Brock, P., Rolnick, D., Donti, P. L. & Kaack, L. Climate Change and AI: Recommendations for Government Action (GPAI, 2021).

  60. Debnath, R., Creutzig, F., Sovacool, B. K. & Shuckburgh, E. Harnessing human and machine intelligence for planetary-level climate action. npj Clim. Action 2, 20 (2023).

    Article  Google Scholar 

  61. Al Khourdajie, A., Skea, J. & Green, R. Climate ambition, background scenario or the model? Attribution of the variance of energy-related indicators in global scenarios. Energy Clim. Change 5, 100126 (2024).

    Article  CAS  Google Scholar 

  62. Li, P.-H., Pye, S., Keppo, I., Jaxa-Rozen, M. & Trutnevyte, E. Revealing effective regional decarbonisation measures to limit global temperature increase in uncertain transition scenarios with machine learning techniques. Clim. Change 176, 80 (2023).

    Article  Google Scholar 

  63. Riahi, K. & Krey, V. Shared Socioeconomic Pathways Scenario Database (SSP). IIASA https://iiasa.ac.at/models-tools-data/ssp (2024).

  64. van Vuuren, D. et al. The Scenario Model Intercomparison Project for CMIP7 (ScenarioMIP-CMIP7). Preprint at https://doi.org/10.5194/egusphere-2024-3765 (2025).

  65. Wong, C. How AI is improving climate forecasts. Nature 628, 710–712 (2024).

    Article  CAS  Google Scholar 

  66. Harder, P. The Role of AI in Responding to Climate Challenges. In Proceedings of AAAI 2022 Fall Symposium (AAAI Press, 2022).

  67. Kikstra, J. S. et al. The IPCC Sixth Assessment Report WGIII climate assessment of mitigation pathways: from emissions to global temperatures. Geosci. Model Dev. 15, 9075–9109 (2022).

    Article  Google Scholar 

  68. Berndt, D. J. & Clifford, J. Using dynamic time warping to find patterns in time series. In Proc. 3rd International Conference on Knowledge Discovery and Data Mining 359–370 (AAAI Press, 1994).

  69. McInnes, L., Healy, J. & Melville, J. UMAP: uniform manifold approximation and projection for dimension reduction. Preprint at https://arxiv.org/abs/1802.03426 (2018).

  70. Lundberg, S. M. & Lee, S.-I. A unified approach to interpreting model predictions. Adv. Neural Inf. Process. Syst. 30, 4768–4777 (2017).

    Google Scholar 

  71. Chen, T. & Guestrin, C. XGBoost: a scalable tree boosting system. In Proc. 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining 785–794 (Association for Computing Machinery, 2016).

  72. Nikitin, A., Iannucci, L. & Kaski, S. TSGM: a flexible framework for generative modeling of synthetic time series. Adv. Neural Inf. Process. Syst. 37, 129042–129061 (2024).

    Google Scholar 

  73. Farahani, A., Voghoei, S., Rasheed, K. & Arabnia, H. R. A brief review of ___domain adaptation. In Advances in Data Science and Information Engineering. Proceedings from ICDATA 2020 and IKE 2020 (eds Stahlbock, R. et al.) 877–894 (Springer, 2021).

  74. Zhu, R. 30,000 Synthetic scenarios in “Using deep learning to generate key variables in global mitigation scenarios”. Zenodo https://doi.org/10.5281/zenodo.15240553 (2025).

  75. Li, P. Anderson111-maker/Deep-IAM: code availablity for NCC (v1.0.0). Zenodo https://doi.org/10.5281/zenodo.15280653 (2025).

Download references

Acknowledgements

This work was supported by the National Key R&D Program of China (grant no. 2024YFF1307000), the National Natural Science Foundation of China (grant no. 72474002), Emerging Engineering Interdisciplinary Young Scholars Project, Peking University, and the Fundamental Research Funds for the Central Universities. H.M. was supported by the National Research Foundation of Korea (grant no. RS-2024-00467678). DL model training was supported by the High-Performance Computing Platform of Peking University. E.B. received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement no. 821471 (ENGAGE).

Author information

Author notes

  1. These authors contributed equally: Peijin Li, Rongqi Zhu.

Authors and Affiliations

  1. Center for Machine Learning Research and Center for Data Science, Peking University, Beijing, China

    Peijin Li & Peijie Zhou

  2. College of Environmental Sciences and Engineering, Peking University, Beijing, China

    Rongqi Zhu & Yang Ou

  3. Graduate School of Green Growth and Sustainability, Korea Advanced Institute of Science and Technology, Daejeon, Republic of Korea

    Haewon McJeon

  4. International Institute for Applied Systems Analysis, Laxenburg, Austria

    Edward Byers

  5. AI for Science Institute, Beijing, China

    Peijie Zhou

  6. National Engineering Laboratory for Big Data Analysis and Applications, Beijing, China

    Peijie Zhou

  7. Institute of Carbon Neutrality, Peking University, Beijing, China

    Yang Ou

Authors
  1. Peijin Li
    View author publications

    Search author on:PubMed Google Scholar

  2. Rongqi Zhu
    View author publications

    Search author on:PubMed Google Scholar

  3. Haewon McJeon
    View author publications

    Search author on:PubMed Google Scholar

  4. Edward Byers
    View author publications

    Search author on:PubMed Google Scholar

  5. Peijie Zhou
    View author publications

    Search author on:PubMed Google Scholar

  6. Yang Ou
    View author publications

    Search author on:PubMed Google Scholar

Contributions

Y.O. and P.Z. jointly designed the study. P.Z. supervised the DL techniques. Y.O. supervised the IAM data analysis. E.B. contributed to modelling data compilation. H.M. contributed to data interpretation. P.L. processed the data and developed the DL models. P.L., R.Z. and Y.O. wrote the paper with input from all co-authors.

Corresponding authors

Correspondence to Peijie Zhou or Yang Ou.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Climate Change thanks Alaa Al Khourdajie, Nikola Milojevic-Dupont and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Distribution of selected key features in 2050 in the AR6 Scenarios Database.

a-e, Boxplots of primary coal consumption (EJ), primary gas consumption (EJ), primary oil consumption (EJ), final liquids consumption (EJ), and carbon sequestration (MtCO2) by aggregated scenario categories of C1234, C56, and C78. f-j, Boxplots of the same variables grouped by the type of IAMs. The central line within each box indicates the median, while the bottom and top edges correspond to the 25th and 75th percentiles (the interquartile range, IQR). The whiskers extend to the furthest data points within 1.5 times the IQR from the lower and upper quartiles.

Source data

Extended Data Fig. 2 Distribution of electricity sector variables in 2050 and 2100 in the AR6 Scenarios Database (blank boxes) and synthetic scenarios (solid boxes).

a-c, Boxplots of electricity generation by fuels by aggregated scenario categories of C1234, C56, and C78 in 2050. d-f, Boxplots of the same variables in 2100. The central line within each box indicates the median, while the bottom and top edges correspond to the 25th and 75th percentiles (the interquartile range, IQR). The whiskers extend to the furthest data points within 1.5 times the IQR from the lower and upper quartiles.

Source data

Extended Data Table 1 Summary of scenarios passed climate vetting by climate category and submission models
Full size table
Extended Data Table 2 Comparison of label transfer performance of different generative deep learning methods in two experimental conditions
Full size table
Extended Data Table 3 Detailed architecture and parameters in the VAE model
Full size table

Supplementary information

Supplementary Information

Supplementary Texts 1–3, Figs. 1–10 and Tables 1–3.

Source data

Source Data Fig. 2

Statistical source data.

Source Data Fig. 3

Statistical source data.

Source Data Fig. 4

Statistical source data.

Source Data Fig. 5

Statistical source data.

Source Data Extended Data Fig. 1

Statistical source data.

Source Data Extended Data Fig. 2

Statistical source data.

Source Data Extended Data Table 1

Statistical source data.

Source Data Extended Data Table 2

Statistical source data.

Source Data Extended Data Table 3

Statistical source data.

Rights and permissions

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.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Li, P., Zhu, R., McJeon, H. et al. Using deep learning to generate key variables in global mitigation scenarios. Nat. Clim. Chang. (2025). https://doi.org/10.1038/s41558-025-02352-8

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1038/s41558-025-02352-8

Search

Advanced search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing