Bayesian estimation of Earth's climate sensitivity and transient climate response from observational warming and heat content datasets

Goodwin, Philip; Cael, B. B.

doi:https://doi.org/10.5194/esd-12-709-2021

Articles | Volume 12, issue 2

https://doi.org/10.5194/esd-12-709-2021

© Author(s) 2021. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/esd-12-709-2021

© Author(s) 2021. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 12, issue 2

Research article

|

17 Jun 2021

Research article |

| 17 Jun 2021

Bayesian estimation of Earth's climate sensitivity and transient climate response from observational warming and heat content datasets

Philip Goodwin and B. B. Cael

Download

Final revised paper (published on 17 Jun 2021)
Supplement to the final revised paper
Preprint (discussion started on 20 Oct 2020)
Supplement to the preprint

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Review of Goodwin and Cael', Anonymous Referee #1, 29 Nov 2020
- AC1: 'Reply to Reviewer #1', P. Goodwin, 10 Dec 2020
RC2: 'Review', Anonymous Referee #2, 30 Nov 2020
- AC2: 'Reply to Reviewer #2', P. Goodwin, 10 Dec 2020

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

ED: Reconsider after major revisions (10 Dec 2020) by Andrey Gritsun

AR by P. Goodwin on behalf of the Authors (26 Mar 2021) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (26 Mar 2021) by Andrey Gritsun

RR by Anonymous Referee #1 (31 Mar 2021)

Suggestions for revision or reasons for rejection

The paper has improved in this iteration. A few additional comments before I can recommend publication: I'm particually interested in a better explanation for the choice of priors, and its influence on the conclusion.

Major comments
* there are a couple of instances where the authors use variable names that deviate from the standard within the literature. For instance, they call their climate sensitivity 'effective climate sensitivity', a term usually reserved for estimates of climate sensitivity for which lambda is assumed constant (so underselling their work). Furthermore, they abbreviate effective climate sensitivity as ECS, which normally stands for 'equilibrium climate sensitivity'. I urge the authors to replace their use of the term ECS with S, with appropriate subscripts (So S_{20 years}, with S_{140} corresponding to ECS). See https://www.ipcc.ch/sr15/chapter/glossary/, under climate sensitivity.
* the prior on lambda multidecadal seems to have a strong influence on the conclusions. The main text refers to the supplementary information for explanation about the prior given, but I don't really see much information there either. I disagree with the other reviewer (sorry! I know how annoying disagreeing reviewers are..) that a prior symmetrical around zero is physically justified. Evidence from climate models consistently show that changes in lambda on that timescale have an amplifying effect. The authors indicate they want an estimate of S independent of model evidence, but section 4.2 of Sherwood also summarizes observational and process-based argumentation for a decreasing lambda.
** I think the previous asymmetry in the prior was fine. If the authors, torn between reviewers, don't want to revert, I'm happy to see the two priors side-by-side.
** Please provide a paragraph of explanation of prior choices in the main text
* the authors use a uniform prior for lambda multidecadal and lambda fast, which is typically not recommended (see the STAN user manual: https://github.com/stan-dev/stan/wiki/Prior-Choice-Recommendations). For consistency, the authors should use the same distribution for lambda Planck, lambda fast and lambda multidecadal (normal distribution seems fine, as the overall feedback is the sum of local feedbacks). There is no physical reason to say that lambda is necessarily between the two given boundaries, and the fact that the posterior does not drop to zero at the boundary indicates the data may indeed by consistent with lambdas < 3.

Minor comments:
multi-diurnal -> multiday?
line 380: incomplete sentence

Hide

ED: Publish subject to minor revisions (review by editor) (14 Apr 2021) by Andrey Gritsun

AR by P. Goodwin on behalf of the Authors (22 Apr 2021) Author's response Author's tracked changes Manuscript

ED: Publish as is (18 May 2021) by Andrey Gritsun

AR by P. Goodwin on behalf of the Authors (19 May 2021) Author's response Manuscript

Download

Article (3205 KB)
Full-text XML

Short summary

Climate sensitivity is a key measure of how sensitive Earth's climate is to human release of greenhouse gasses, such as from fossil fuels. However, there is still uncertainty as to the value of climate sensitivity, in part because different climate feedbacks operate over multiple timescales. This study assesses hundreds of millions of climate simulations against historical observations to reduce uncertainty in climate sensitivity and future climate warming.