Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Feb 7;120(6):e2207183120.
doi: 10.1073/pnas.2207183120. Epub 2023 Jan 30.

Data-driven predictions of the time remaining until critical global warming thresholds are reached

Noah S Diffenbaugh  1 , Elizabeth A Barnes  2
Affiliations

Data-driven predictions of the time remaining until critical global warming thresholds are reached

Noah S Diffenbaugh et al. Proc Natl Acad Sci U S A. .

Abstract

Leveraging artificial neural networks (ANNs) trained on climate model output, we use the spatial pattern of historical temperature observations to predict the time until critical global warming thresholds are reached. Although no observations are used during the training, validation, or testing, the ANNs accurately predict the timing of historical global warming from maps of historical annual temperature. The central estimate for the 1.5 °C global warming threshold is between 2033 and 2035, including a ±1σ range of 2028 to 2039 in the Intermediate (SSP2-4.5) climate forcing scenario, consistent with previous assessments. However, our data-driven approach also suggests a substantial probability of exceeding the 2 °C threshold even in the Low (SSP1-2.6) climate forcing scenario. While there are limitations to our approach, our results suggest a higher likelihood of reaching 2 °C in the Low scenario than indicated in some previous assessments-though the possibility that 2 °C could be avoided is not ruled out. Explainable AI methods reveal that the ANNs focus on particular geographic regions to predict the time until the global threshold is reached. Our framework provides a unique, data-driven approach for quantifying the signal of climate change in historical observations and for constraining the uncertainty in climate model projections. Given the substantial existing evidence of accelerating risks to natural and human systems at 1.5 °C and 2 °C, our results provide further evidence for high-impact climate change over the next three decades.

Keywords: AI for climate; CMIP6; UN Paris agreement; global warming; machine learning.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interest.

Figures

Fig. 1.
Fig. 1.
Time to global warming thresholds in global climate model ensembles. (A) Global temperature change relative to the preindustrial baseline (1850 to 1899) for 10-member global climate model ensembles in the High (SSP3-7.0), Intermediate (SSP2-4.5) and Low (SSP1-2.6) climate forcing scenarios. Gray lines show individual realizations; colors show the mean of the respective 10 realizations for each global climate model. See SI Appendix, Table S1 for list of climate models used in each climate forcing scenario. (B) Maps of temperature anomalies for the “threshold year” (i.e., the year in which the ensemble-mean global warming reaches 1.5 °C) for the global climate models with the earliest and latest threshold years in SSP3-7.0. Anomalies are shown relative to the 1951 to 1980 baseline to match the baseline period of the temperature observations (see Materials and Methods). (C) Comparison of training, validation, and testing of the artificial neural network (ANN) trained on maps of annual temperature and a global warming threshold of 1.5 °C in SSP3-7.0. Left panel shows the predicted number of years until the 1.5 °C threshold for each annual temperature map in each global climate model. Right panel shows the comparison of training, validation, and testing for the predicted versus true number of years until the 1.5 °C threshold across the full global climate model ensemble (SI Appendix, Table S1). See SI Appendix, Figs. S1–S3 for additional temperature thresholds and scenarios.
Fig. 2.
Fig. 2.
Time to the current level of global warming predicted from observed maps of annual temperature anomalies. (A) Maps of observed annual temperature anomalies for selected individual years, including the first year of our observations-based prediction (1980), the year following the Pinatubo volcanic eruption (1992), the year with the highest global-mean temperature (2016), and the most recent year for which annual data are available (2021). (B) The time to 1.1 °C of global warming predicted from the observed map of annual temperature anomalies, using the artificial neural network (ANN) trained on a global warming threshold of 1.1 °C in the High climate forcing scenario (SSP3-7.0). Left panel shows the median prediction (and ±1σ range) for the observed map of annual temperature anomalies in each year from 1970 to 2021. The slope quantifies the rate of change of predicted time to 1.1 °C (with a perfect prediction exhibiting a slope of −1 y per year). The Right panel shows the distribution of predicted years in which 1.1 °C will be reached based on the observed map of annual temperature anomalies in 2021. Note that no historical temperature observations are used in training, validating, or testing the ANN.
Fig. 3.
Fig. 3.
Time to future global warming thresholds predicted from observed maps of annual temperature anomalies. As in Fig. 2C, but for the 1.5 °C and 2 °C thresholds in the High (SSP3-7.0), Intermediate (SSP2-4.5), and Low (SSP1-2.6) climate forcing scenarios. The global climate models used for each scenario/threshold combination are shown in SI Appendix, Table S1.
Fig. 4.
Fig. 4.
Attribution heatmap of the most relevant regions for the artificial neural network (ANN) prediction of the time-to-threshold. (A) Global climate model ensemble. (B) Historical temperature observations for 2018 to 2021. Warm colors indicate shorter time to global warming threshold; cool colors indicate longer time to global warming threshold. Maps show results for the 1.5 °C threshold in the High (SSP3-7.0) climate forcing scenario. See Materials and Methods for additional details of the heatmap calculation and SI Appendix, Figs. S12 and S13 for additional scenarios.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. UNFCCC Adoption of the Paris Agreement. I: Proposal by the President. Draft Decision CP.21 https://unfccc.int/resource/docs/2015/cop21/eng/l09r01.pdf (United Nations Office, Geneva, 2015)
    1. Allen M., Babiker M., Chen Y., de Coninck H. C., “IPCC SR15: Summary for policymakers” in IPCC Special Report Global Warming of 1.5 °C (Intergovernmental Panel on Climate Change, 2018).
    1. IPCC, “Summary for policymakers” in Climate Change 2014: Impacts, Adaptation, and Vulnerability in Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Field C. B., et al., Eds. (Cambridge University Press, 2014), pp. 1–32.
    1. IPCC, “Summary for policy makers” in Climate Change 2022: Impacts, Adaptation and Vulnerability in Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change, Pörtner H. O., et al., Eds. (IPCC, 2022).
    1. Schleussner C.-F., et al. , Differential climate impacts for policy-relevant limits to global warming: The case of 1.5 °C and 2 °C. Earth Syst. Dyn. 7, 327–351 (2016).

LinkOut - more resources