the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 10 Feb 2025
AR6 updates to RF by GHGs and aerosols lowers the probability of accomplishing the Paris Agreement compared to AR5 formulations
Abstract. Many Reduced complexity climate models (RCMs) and Earth System Models (ESMs) use prescribed concentrations or Effective Radiative Forcing (ERF) of Greenhouse Gases (GHGs) and tropospheric aerosols as inputs for projections. Revisions to these datasets, made in Chapter 7 and Annex III of the Sixth IPCC Assessment Report: The Physical Science Basis (AR6, 2021) are vital to ensure the accuracy of climate model forecasts. AR6 provided updates to the formulation of ERF for most GHGs and tropospheric aerosols, relative to values in AR5 (2013). In this work, we provide a comprehensive assessment of how the changes to the ERF datasets impact projections of future warming, using our multiple linear regression energy balance RCM, the Empirical Model of Global Climate (EM-GC). We provide an analysis of the rate of human−induced warming (AAWR) between 1974 and 2014, and Effective Climate Sensitivity (EffCS) from the regression to the observation-based historical climate record with ERF datasets predating the AR6 report (which we term Baseline Framework) and AR6 ERF data (AR6 Framework). Probabilistic projections on future warming that consider the uncertainty in the magnitude of climate feedback and ERF from tropospheric aerosols are provided for four policy-relevant Shared Socioeconomic Pathway (SSP) scenarios. We find AAWR within the AR6 Framework to be 0.18 [0.13 to 0.21 °C decade−1, 5−95 % range], a slight increase to the values of 0.16 [0.12 to 0.20 °C decade−1] within the Baseline Framework. The central estimate of EffCS is found to be nearly identical between the two Frameworks, but a narrower range is found for the AR6 Framework at 2.29 [1.54 to 3.11 °C, 5−95 % range] relative to 2.26 [1.45 to 4.37 °C] within the Baseline Framework. We find Equilibrium Climate Sensitivity (ECS) to be 3.24 [1.92 to 5.15 °C] for the AR6 best estimate of the pattern effect. Our estimates of AAWR, EffCS and ECS are highly consistent with recent studies and observationally constrained CMIP6 model output. Projections of future warming for the AR6 Framework compared to the Baseline Framework show an increase of 0.2 to 0.4 °C in the end-of century median warming for the SSP scenarios studied. This increase corresponds to a significantly lowered possibility of accomplishing the goals of the Paris Agreement (PA). In particular, the SSP2−4.5 scenario, that is widely considered to be consistent with current climate policies, only offers an 8 % chance of accomplishing the PA upper limit of 2.0 °C warming by the end of the century within the AR6 Framework.
- Preprint
(2008 KB) - Metadata XML
-
Supplement
(1033 KB) - BibTeX
- EndNote
Status: open (until 25 Apr 2025)
-
RC1: 'Comment on egusphere-2025-342', Anonymous Referee #1, 31 Mar 2025
reply
Review of AR6 updates to RF by GHGs and aerosols lowers the probability of accomplishing the Paris Agreement compared to AR5 formulations by Farago et al.
Summary
The authors apply an existing multiple linear regression model to decompose the relative contribution of internal and external forcing factors to global mean surface temperature (GMST) change over the 20th and 21st century. They compare the influence of different assumptions about effective radiative forcing from various constituents (but primarily tropospheric aerosols) between two generations of the CMIP protocol (AR5, referred to as Baseline, and AR6). The authors show that their MLR model reproduces the majority of features of the GMST response, and the effective climate sensitivity, simulated by a range of previous simplified and comprehensive modelling efforts. They use this information to provide probabilistic estimates of GMST remaining below Paris targets (1.5C and 2.0C).Major comments
The paper is well researched, and well written; the figures are clear and communicate the main findings of the analysis. I believe that the conclusions reached are appropriate based on the methods and evidence presented. However, I cannot recommend publication of this manuscript for the following reasons.First, this reviewer found that the authors have not adequately communicated what are the primary novel contributions of the research. On the contrary, in virtually all cases in the results sections, the authors highlight that their results are consistent with previous studies. This holds true for previous studies using simplified or intermediate-complexity models, and comprehensive modelling like CMIP6. This is a very well-studied field over the past decade, and the authors must articulate clearly how this research advances the discipline beyond what the myriad of previous studies has done.
Second, while the authors must be commended on their attention to detail and the depth of the research undertaken, the manuscript is much too long considering the paucity of new results being presented. At times it felt like a PhD thesis; for example, Section 2.3 provides many textbook-level definitions of ERF for various atmospheric constituents, and Section 2.5 describes every assumption for the inputs of the regression model in intricate detail. This raised the question of who exactly is the target audience for this work? Since the aim is to publish in ESD, there should be an assumption that interested readers will have sufficient background knowledge in climate change science to trust that emissions inventories, sources of natural/internal climate variability etc. are properly referenced and incorporated without the need for such a detailed assessment here. Potentially important and relevant previous literature was also excluded; for example, see: https://iopscience.iop.org/article/10.1088/1748-9326/6/4/044022/meta.
Third, the linear modelling framework itself does not appear to be a new contribution (e.g., McBride et al. 2021). Therefore, it is somewhat surprising that potentially important limitations of the linear model approach are not investigated or advanced in this research. For example, on L513 the authors describe needing to include a third constraint (consistency with recent observed temperature trends) in order to yield solutions that match the GMST time series over the recent past. Given the importance of ERF_CO2 and ERF_AER on the GMST predictions from the EM-GC model, this suggests a probable role for nonlinear interactions between aerosol and CO2 forcing that the current model cannot capture. It would have been interesting to see this commented on, if not addressed, in the research.
Fourth, one of the major findings of the research highlighted by the authors is the apparent increase in warming rate under the AR6 assumptions compared to pre-AR6 (baseline). However, on L502 the authors state that the 6\% higher climate sensitivity in AR6 comes from applying the published formula for ERF_CO2 from AR6 that is larger than the pre-AR6 formula provided by Myhre et al. (1998). Therefore, it appears that the increased warming rate is "baked-in" to the EM-GC model, rather than an emergent property, making the findings of more warming and a lower probability of remaining below the Paris targets largely unsurprising.
Minor comments:
-L23: The overlap of the baseline and AR6 confidence intervals suggests that the statistical evidence for an increase in the mean is rather weak.-L119 and L122: do the authors mean to say adopted, rather than adapted?
-L152: Can the authors provide the proportions for the different effects in this attribution? Are CO2 concentrations the dominant effect?
-L205: Can the authors comment on why the ERF_AER value changes by so much (15% larger) when the time period is shortened by only 5 years (ending in 2014)?
-L235: This paragraph is very unclear. What is the single best estimate ERF_AER time series? What portion of the difference is highlighted?
-L690: Why did the authors elect to not examine a business-as-usual/ high emission scenario like SSP5-8.5?
-L855: This conclusion is challenging, because the agreement between EM-GC outputs and the observed GMST timeseries is explicitly built in to the EM-GC model, whereas for the majority of CMIP6 models they are freely running through the 20th Century. Whether this reduces the value of those simulations is a matter for debate; perhaps a more nuanced view is that it affects the types of questions that one should ask of the CMIP-class models.
Citation: https://doi.org/10.5194/egusphere-2025-342-RC1 -
AC2: 'Reply on RC1', Endre Farago, 11 Apr 2025
reply
Thank you this review. We have replied to this review, as well as the other review, in the RC2 box because of the overlap in the content of both reviews.
Citation: https://doi.org/10.5194/egusphere-2025-342-AC2
-
AC2: 'Reply on RC1', Endre Farago, 11 Apr 2025
reply
-
RC2: 'Comment on egusphere-2025-342', Anonymous Referee #2, 08 Apr 2025
reply
In this manuscript, the authors use the Empirical Model of Global Change (EM-GC) to explore how updates to effective radiative forcing, as reported in the IPCC Sixth Assessment Report (AR6), influence the following climate metrics: effective climate sensitivity, the rate of attributable anthropogenic warming, and projections of future warming. The manuscript concludes with an assessment of how these updates affect the likelihood of meeting the climate policy goals set by the 2015 Paris Agreement.
I was genuinely excited to see the EM-GC used in this context, and I believe this study has the potential to make a meaningful contribution to the literature. The research question is both timely and important, and I encourage the authors to continue developing this work. However, in its current form, the manuscript faces some structural challenges and clarity issues that make it difficult to fully appreciate the significance of the results. I would strongly encourage the authors to revise and resubmit, as I believe that with improvements, this paper could become a valuable addition to the field.
While I did not feel that a detailed, line-by-line review was appropriate at this stage, I would like to share a few broader comments that I hope will be helpful in guiding the revision.
At times, the manuscript reads like a blend of two distinct papers — part literature review, part research article. I recognize and appreciate the substantial effort the authors have put into the background material, and the breadth of the literature covered is impressive. That said, I feel that the extensive background somewhat overshadows the more novel and exciting aspects of the authors' analysis. I would recommend streamlining the background, particularly in the methods and data sections, and relying more heavily on citations to established work, which would allow the new contributions to stand out more clearly.
I think referencing prior EM-GC literature more explicitly could help improve clarity in the model description. For example, Section 2.4 closely resembles McBride et al. (2021), and equations (1)–(4) appear to be the same as those in that manuscript. Could the authors clarify whether these equations are indeed unchanged, or if they have been modified in this study? Providing that clarification will help situate the current work within the existing EM-GC framework and highlight any new developments more effectively.
Another question that arose when reading this manuscript was how do the authors deal with the issue with the ERF of aerosols in the AR6 analysis (see Zelinka 2023). Did the authors account for or correct for this basis their analysis? A comment of if/how the authors address this should be included in the manuscript or discussed as a potential limitation of the study.
I appreciate the authors’ effort and I am confident that with thoughtful revisions, this work can make a valuable impact. I look forward to seeing a future version of this manuscript and the contributions it will bring to our field.
Citation: https://doi.org/10.5194/egusphere-2025-342-RC2 -
AC1: 'Reply on RC2', Endre Farago, 11 Apr 2025
reply
We thank both Reviewers for taking the time to carefully read our manuscript and provide valuable comments. We now understand the Reviewers’ concerns regarding the length of the manuscript. if allowed to proceed, we propose to revise the manuscript by greatly reduce the length, as described below.
The revised manuscript would be built around four key figures:
- (New Figure 1): Comparison of AAWR and EffCS between the Baseline and AR6 Simulations (Original Fig. 3).
- (New Figure 2): Comparison of Equilibrium Climate Sensitivity (ECS) as the function of the pattern effect parameter α’ (Original Fig. 5).
- (New Figure 3): Comparison of time-dependent probabilistic projections for the four SSP scenarios studied (Original Fig. 7).
- (New Figure 4): Comparison of Probability Distribution Functions (PDFs) for the end-of-century warming for the four SSP scenarios studied (Original Fig. 8).
The figures would be reduced to four, relative to the eight figures in the main body of the initial manuscript. Additionally, we propose moving Figs. 1 and 6 of the original manuscript to the supplement, and completely removing Figs. 2 and 4 of the original manuscript.
We also propose to revise and greatly shorten the text of the manuscript to better reflect the key results of our paper, given below, which we now realize was obscured by the length of the introductory material:
- Projected ERF provided in AR6 for the SSPs is much greater than in the prior SSP dataset, from the original (baseline) SSP database. This excess is due mainly to updates in the ERF formulations for CO2 and CH4. Further, for each SSP scenario, the projected ERF given by AR6 at the end of this century significantly exceeds the target radiative forcing associated with each given SSP scenario (original Fig 1, to be moved to Supplement).
- Historical GMST, when fit with the AR6 ERF datasets, corresponds to a narrower range of EffCS of 2.29 ⁰C [1.54 ⁰C to 3.11 ⁰C] relative to EffCS inferred from the Baseline simulations 2.26 ⁰C [1.45 ⁰C to 4.37 ⁰C]. (New Fig. 1)
- We provide new estimates of ECS using various values of the pattern effect parameter α’, and find a range for ECS of 3.24 [1.92 to 5.15 ⁰C] for the AR6 best estimate of α’ (New Fig. 2). This analysis is a notable advance relative to the McBride et al. (2021) paper.
- When the AR6 ERF datasets are used, the simulated GMST in the future is considerably higher than that for Baseline simulations, a direct consequence of the increase in projected ERF, resulting in a less optimistic chance of achieving the 1.5C and 2.0C goal and upper limit of the Paris Agreement (New Fig. 4)
- Our AR6-based forecasts of GMST still provide a lower projected warming than is given by many of the CMIP6 ensemble members (New Fig. 4). We propose to update Fig. 7 of the original manuscript, such that we shall now include the minimum-, maximum and mean projections of time-dependent GMST projections from CMIP6, as our New Fig. 3.
Again, we sincerely appreciate both reviews and we hope we will be allowed to submit a revised, significantly shortened manuscript in response to these comments. The revised manuscript would rely heavily on the McBride et al. (2021) paper for our methodology, as suggested by both Reviewers.
Citation: https://doi.org/10.5194/egusphere-2025-342-AC1
-
AC1: 'Reply on RC2', Endre Farago, 11 Apr 2025
reply
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 147 | 50 | 12 | 209 | 19 | 8 | 11 |
- HTML: 147
- PDF: 50
- XML: 12
- Total: 209
- Supplement: 19
- BibTeX: 8
- EndNote: 11
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 84 | 43 |
| China | 2 | 18 | 9 |
| Germany | 3 | 12 | 6 |
| United Kingdom | 4 | 8 | 4 |
| undefined | 5 | 7 | 3 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 84