Halving of the uncertainty in projected warming over the past decade | npj Climate and Atmospheric Science - Nature.com

Halving of the uncertainty in projected warming over the past decade | npj Climate and Atmospheric Science

Uncertainties in projected 21st century warming were very large a decade ago, increasing the costs of climate change adaptation, especially those associated with long-lived infrastructure. Here we show that through progress in climate policy and climate science, these uncertainties have decreased dramatically over the past decade.

Ten years ago, the Intergovernmental Panel on Climate Change (IPCC) published projections of 2081–2100 warming which ranged from 0.9 °C to 5.4 °C relative to preindustrial 1(Fig. 1a). Since many climate changes and climate impacts scale with global warming 2, 3, the large uncertainty in this projection translates into similarly large uncertainties in projections of many key climate impacts. For an engineer designing a bridge, a forester planting trees, or a city planning coastal defenses, adaptation to a climate with global warming of 5.4 °C versus adaptation to a climate with global warming of 0.9 °C would look very different. Hence this large uncertainty in projected global warming can impose large real costs associated with insufficient adaptation, or unnecessarily stringent adaptation 4, 5. How has this uncertainty evolved over the past ten years?

Panel (a) shows distributions of projected global mean warming in 2081–2100 relative to 1850–1900 from the IPCC AR5 1, based on assessed likely ranges of projected warming relative to 1986–2005 in the four Representative Concentration Pathway (RCP) scenarios 10 shown and the assessed 90% confidence interval of observed warming in 1986–2005 relative to 1850–1900, assuming uncorrelated Gaussian uncertainties. Similarly, panel (b) shows corresponding distributions of warming under three Shared Socioeconomic Pathway (SSP) based scenarios 6 from the IPCC AR6 13, based on assessed very likely ranges, which were estimated using observational constraints. Heavy black lines in (a, b) show average distributions over each set of scenarios, based on an assumption of equal likelihood for each of the scenarios shown, and uncertainty bars are the corresponding 90% confidence intervals. Panel (c) shows CO2-equivalent greenhouse gas emissions in the five SSP-based scenarios considered in IPCC AR6 12, and as observed in 2019, calculated based on 100-year Global Warming Potential. The dark grey line shows the corresponding CO2-equivalent emissions in the current policies scenario in the 2023 UNEP Emissions Gap Report (EGR) 17, and the shaded grey range shows the full range of future emissions considered in that report.

Uncertainties in projected warming stem from three sources: uncertainty in future emissions, uncertainty in the climate response to those emissions, and internal variability 4. Uncertainty in future emissions is typically explored by starting with storylines of possible future socio-economic and climate policy changes 6, but more recently has been investigated by probabilistically modelling future changes in these factors 7, 8, 9; uncertainty in the climate response may be estimated simply from the spread across a set of climate models, or may be constrained using observations of past climate or climate change 2; internal variability is irreducible in the long term, but can be constrained in the near term using decadal projections 4. In 2009, Hawkins and Sutton

Narrowing of emissions uncertainty

The IPCC Fifth Assessment Report (AR5) projections 1 were based on four scenarios spanning a range of radiative forcing from approximately 2.6 Wm−2 to 8.5 Wm−2 in 2100 10. These scenarios, developed before the Paris Agreement, spanned a range of futures from ongoing rapid fossil fuel development, through to stringent climate action 1, but those using these projections for planning climate change adaptation had to make some assumption concerning their relative likelihood, and often used all the scenarios with equal weight, or the lowest and highest scenarios to bracket uncertainties 11. In Fig. 1a we show the result of assuming that the scenarios are equally probable (black distribution), which results in a 5–95% confidence interval of projected warming in 2081–2100 of 0.9–5.3 °C.

Since the publication of the IPCC AR5 1, adoption of the Paris Agreement and subsequent agreements including the Glasgow Climate Pact reflect real progress in international commitments to reduce emissions. Although global emissions are still increasing, they are doing so at a slower rate than over earlier decades, and at least 18 countries have seen sustained reductions in greenhouse gas emissions over more than a decade, while the costs of low emissions technologies are falling rapidly 12. While the Working Group I component of the IPCC Sixth Assessment Report (AR6) showed global warming projections for scenarios with an even wider range of radiative forcings in 2100 than AR5, from 1.9 Wm−2 to 8.5 Wm−2 (Fig. 1c) 2, 13, a discussion of the likelihood of the scenarios concluded that the most fossil fuel intensive scenario considered, SSP5-8.5, was unlikely, that SSP5-8.5 and SSP3-7.0 should be considered counter-factual scenarios, and that current policies would lead to approximately stable greenhouse gas emissions similar to those in SSP2-4.5 14. This conclusion is supported by other recent studies which find that scenarios with substantial increases in emissions are unlikely 7, 9, 15, 16. Moreover, the United Nations Environment Programme (UNEP) 2023 Emissions Gap Report 17 recently showed a scenario of future greenhouse gas emissions based on current policies with median 2050 greenhouse gas emissions close to those of SSP2-4.5 (Fig. 1c). If countries deliver on emissions reductions by 2030 pledged under the Paris Agreement (so-called Nationally Determined Contributions), then 2050 emissions are projected to be substantially lower, and lower still if countries also deliver on their pledged zero emissions targets – though probably not low enough to limit global warming to 1.5 °C 17. The full range of scenarios considered in the Emissions Gap Report 6, 14, to the current policies scenario (shown as the grey shaded region in Fig. 1c). The narrowing of emissions uncertainty due to climate policy developments has been augmented by developments in probabilistic scenario modelling, which have also helped constrain emissions uncertainty 7, 8, 9.

Given these considerations, the range of plausible future emissions is now lower and narrower than it was in 2013 at the time of publication of the IPCC AR5 16, and can reasonably be bracketed by the SSP1-1.9 and SSP2-4.5 scenarios (Fig. 1b, c). We note that this set of three SSP-based scenarios does not cover the full range of possible non-CO2 greenhouse gas and aerosol emissions trajectories, but we do not expect the range of 2081–2100 global mean warming to be strongly sensitive to this, given that low greenhouse emissions in the strongly mitigated scenarios are unlikely to be compatible with high emissions of cooling aerosols, and that current climate policies limit emissions of a basket of greenhouse gas emissions, not just CO2. This set of scenarios also does not sample over possible external shocks, through recent examples such as the COVID-19 pandemic and the war in Ukraine have had only a limited effect on long-term greenhouse gas emissions 18, 19. It is of course still possible that governments could backslide on climate action and future emissions could be higher than implied by the Emissions Gap Report current policies scenario 17, but given existing political commitments and technological changes this seems unlikely. Moreover, other possible but unlikely eventualities could alter radiative forcing and global temperatures in 2081–2100, such as a large volcanic eruption 2, geoengineering by stratospheric aerosol injection 20, but such eventualities are not typically accounted for in climate projections or adaptation planning 2.

Narrowing of climate response uncertainty

At the same time as emissions uncertainty has narrowed, climate response uncertainty has also narrowed very substantially. Projections of warming in the IPCC AR5 1 were made relative to a 1986–2005 baseline by taking the 90% range of simulated warming in available climate model simulations and assessing this as a 66% (likely) confidence interval, to account for uncertainties not sampled over in the simulations. Since the publication of the IPCC AR5, the use of observational constraints on climate warming and climate sensitivity has advanced very substantially. As first suggested more than twenty years ago 21 observations of past warming or other climate variables can be used to constrain climate model projections of future climate change. If an observable property of the climate system, such as the past warming trend or a variable related to climate feedbacks, is correlated with projected warming across models, this relationship, together with observations of the variable concerned, can be used to constrain projected warming 2. Recent literature on this topic using either the past global warming trend to constrain projected warming or other observables to constrain climate sensitivity was assessed by the IPCC AR6 and used to derive an observationally-constrained range of projected future warming under each scenario that was considerably narrower than the raw model-simulated range 13(Fig. 1b). The lengthening observational record has also meant that a larger fraction of 2081–2100 warming relative to 1850–1900 can be directly observed, rather than estimated from models, further reducing uncertainties. Overall, these factors contributed to a very substantial narrowing of the range of assessed warming for individual scenarios in IPCC AR6 compared to AR5: the projected ranges of warming under SSP1-2.6 and SSP2-4.5 (Fig. 1a). The mean warming projected under SSP1-2.6 and SSP2-4.5 in AR6 (Fig. 1b) is also higher compared to warming projected under RCP2.6 and RCP4.5 in AR5 (Fig. 1a). This is explainable in part by a higher radiative forcing in the nominally equivalent SSP scenarios compared to corresponding RCP scenarios 2, as well as by upward revisions to the baseline observed warming 22.

Synthesis

Like the IPCC AR5, the AR6 also did not assign probabilities to individual emissions scenarios, beyond assessing that the likelihood of SSP5-8.5 was low 14. However, if we make the simplifying assumption of assigning equal probabilities to each of the three SSP-based scenarios consistent with the range bracketed by the UNEP Emissions Gap Report 2050 emissions 17(Fig. 1c), we obtain the black distribution of projected warming in 2081–2100 shown in Fig. 1b, which has a 90% confidence interval of 1.1–3.2 °C, which is less than half of the width of the corresponding range of projected warming based on the IPCC AR5 (Fig. 1a). Alternatively, we can assign weights to the SSP-based scenarios based on a published probabilistic scenario analysis 9. Weighting the SSP-based scenarios in order to minimize the root mean square difference between their cumulative frequency distribution of CO2 emissions in 2050 and that of those of the five weighted scenario clusters from the probabilistic scenario analysis 1a).

Adaptation costs are estimated to be $215bn per year this decade in developing countries 23 and more on a global basis, and uncertainties in climate projections contribute to these costs, especially for long-lived infrastructure 4, 5. The dramatic narrowing of uncertainties in projected warming described here, which is a direct result of progress in climate policy and climate science, therefore has large potential value for long-term planning of climate change adaptation. Nonetheless, a range of end-of-century warming from approximately the current level to about three times that level, which would be associated with very high risks and impacts, is still a very large uncertainty range. Hence, there is still much work to be done by policymakers to strengthen climate action to meet the Paris Agreement goals 12, 17, as well as by climate scientists to further reduce climate response uncertainty 4.

Data availability

Data shown in Fig. 1 were taken directly from the IPCC AR5 12, 13, and the 2023 UNEP Emissions Gap Report 17. Data used to calculate the projected warming based on a published probabilistic scenario analysis were taken from Moore et al. 9(source data for Fig 3; https://static-content.springer.com/esm/art%3A10.1038%2Fs41586-022-0442…) and IPCC 2021 13(source data for Fig. SPM.4; https://dap.ceda.ac.uk/badc/ar6_wg1/data/spm/spm_04/v20210809/panel_a/C…).

1.
IPCC Summary for Policymakers. In Climate Change 2013: The Physical Science Basis (eds Stocker, T.F. et al.) 3–29 (Cambridge Univ. Press, 2013).
2.
Lee, J.-Y. et al. Future global climate: Scenario-based projections and near-term information. Chapter 4. In Climate Change 2021: The Physical Science Basis, 553–672 (Cambridge University Press, 2021).
3.
IPCC Summary for Policymakers. In Climate Change 2022: Impacts, Adaptation and Vulnerability (eds Pörtner, H.-O. et al.) 3–33 (Cambridge Univ. Press, 2022).
4.
Hawkins, E. & Sutton, R. The potential to narrow uncertainty in regional climate predictions. Bull. Am. Meteorol. Soc. 90, 1095–1108 (2009).
5.
Eisenack, K. & Paschen, M. Adapting long-lived investments under climate change uncertainty. J. Environ. Econ. Manag. 116, 102743 (2022).
6.
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).
7.
Liu, P. R. & Raftery, A. E. Country-based rate of emissions reductions should increase by 80% beyond nationally determined contributions to meet the 2 °C target. Commun. Earth Environ. 2, 29 (2021).
8.
Huard, D., Fyke, J., Capellán‐Pérez, I., Matthews, H. D. & Partanen, A.-I. Estimating the likelihood of GHG concentration scenarios from probabilistic Integrated Assessment Model simulations. Earth’s Future 10, e2022EF002715 (2022).
9.
Moore, F. C. et al. Determinants of emissions pathways in the coupled climate–social system. Nature 603, 103–111 (2022).
10.
Van Vuuren, D. P. et al. The representative concentration pathways: an overview. Clim. Change 109, 5–31 (2011).
11.
Ho, E., Budescu, D. V., Bosetti, V., van Vuuren, D. P. & Keller, K. Not all carbon dioxide emission scenarios are equally likely: A subjective expert assessment. Clim. Change 155, 545–561 (2019).
12.
IPCC Summary for Policymakers. In Climate Change 2022: Mitigation of Climate Change (eds Shukla, P. R. et al.) 3–48 (Cambridge University Press, 2022).
13.
IPCC Summary for Policymakers. In Climate Change 2021: The Physical Science Basis (eds Masson-Delmotte, V. et al.) 3–32 (Cambridge University Press, 2021).
14.
Chen, D. et al. Framing, context, and methods. Chapter 1. In Climate Change 2021: The Physical Science Basis, 147–286 (Cambridge University Press, 2021).
15.
Hausfather, Z. & Peters, G. P. Emissions – the ‘business as usual’ story is misleading. Nature 577, 618–620 (2020).
16.
Lehner, F., Hawkins, E., Sutton, R., Pendergrass, A. G. & Moore, F. C. New potential to reduce uncertainty in regional climate projections by combining physical and socio‐economic constraints. AGU Adv. 4, e2023AV000887 (2023).
17.
UNEP Emissions Gap Report 2023 (UNEP, 2023).
18.
Ray, R. L., Singh, V. P., Singh, S. K., Acharya, B. S. & He, Y. What is the impact of COVID-19 pandemic on global carbon emissions? Sci. Total Environ. 816, 151503 (2022).
19.
van Meijl, H. et al. The Russia-Ukraine war decreases food affordability but could reduce global greenhouse gas emissions. Commun. Earth Environ. 5, 59 (2024).
20.
Coupe, J., Bardeen, C. G., Robock, A. & Toon, O. B. Nuclear winter responses to nuclear war between the United States and Russia in the Whole Atmosphere Community Climate Model Version 4 and the Goddard Institute for Space Studies ModelE. J. Geophys. Res. Atmos. 124, 8522–8543 (2019).
21.
Allen, M. R., Stott, P. A., Mitchell, J. F. B., Schnur, R. & Delworth, T. L. Quantifying the uncertainty in forecasts of anthropogenic climate change. Nature 407, 617–620 (2000).
22.
Gulev, S. K. et al. Changing state of the climate system. Chapter 2. In Climate Change 2021: The Physical Science Basis, 287–422 (Cambridge University Press, 2021).
23.
UNEP Adaptation Gap Report 2023 (UNEP, 2023).

Acknowledgements

N.P.G. acknowledges helpful comments on the manuscript from Elizabeth Bush (Environment and Climate Change Canada).

Ethics declarations

Competing interests

The author declares no competing interests.

Additional information

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

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

About this article

Cite this article

Gillett, N.P. Halving of the uncertainty in projected warming over the past decade. npj Clim Atmos Sci 7, 146 (2024). https://doi.org/10.1038/s41612-024-00693-3

-
Received:
-
Accepted:
-
Published:
-
DOI: https://doi.org/10.1038/s41612-024-00693-3