Tropospheric ozone in CCMI models and Gaussian process emulation to understand biases in the SOCOLv3 chemistry–climate model

Revell, Laura E.; Stenke, Andrea; Tummon, Fiona; Feinberg, Aryeh; Rozanov, Eugene; Peter, Thomas; Abraham, N. Luke; Akiyoshi, Hideharu; Archibald, Alexander T.; Butchart, Neal; Deushi, Makoto; Jöckel, Patrick; Kinnison, Douglas; Michou, Martine; Morgenstern, Olaf; O'Connor, Fiona M.; Oman, Luke D.; Pitari, Giovanni; Plummer, David A.; Schofield, Robyn; Stone, Kane; Tilmes, Simone; Visioni, Daniele; Yamashita, Yousuke; Zeng, Guang

doi:https://doi.org/10.5194/acp-18-16155-2018

Articles | Volume 18, issue 21

Atmos. Chem. Phys., 18, 16155–16172, 2018

https://doi.org/10.5194/acp-18-16155-2018

© Author(s) 2018. This work is distributed under
the Creative Commons Attribution 4.0 License.

Special issue: Chemistry–Climate Modelling Initiative (CCMI) (ACP/AMT/ESSD/GMD...

Atmos. Chem. Phys., 18, 16155–16172, 2018

https://doi.org/10.5194/acp-18-16155-2018

© Author(s) 2018. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 18, issue 21

Research article 13 Nov 2018

Research article | 13 Nov 2018

Tropospheric ozone in CCMI models and Gaussian process emulation to understand biases in the SOCOLv3 chemistry–climate model

Laura E. Revell et al.

Download

Final revised paper (published on 13 Nov 2018)
Supplement to the final revised paper
Preprint (discussion started on 26 Jun 2018)
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 "Tropospheric ozoneo in CCMI and Gaussian emulation to understand biases in the SOCOL v3 chemistry-climate model."', Edmund Ryan, 01 Aug 2018
- AC1: 'Response to E Ryan', Laura Revell, 12 Sep 2018
RC2: 'Review of Revell et al.', Anonymous Referee #2, 25 Aug 2018
- AC2: 'Response to Reviewer 2', Laura Revell, 12 Sep 2018

Peer-review completion

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

AR by Laura Revell on behalf of the Authors (12 Sep 2018) Author's response Manuscript

ED: Referee Nomination & Report Request started (12 Sep 2018) by Paul Young

RR by Anonymous Referee #2 (24 Sep 2018)

RR by Edmund Ryan (25 Sep 2018)

Suggestions for revision or reasons for rejection

Thank-you for thoroughly addressing all of my comments. The paper looks great and the description of the emulator + sensitivity analysis parts are much clearer now. Just four really quick things it'd be great if you could change also:
(1) Please change the first sentence of section 2.4 to: "Variance based global sensitivity analysis quantifies the contribution of a single parameter to the variance of a model's output."
(2) Figure 3. The error bars seem quite large here for a few of the validation points. What have you used for the "emulator uncertainty"? If it's the emulator variance then this is incorrect since the variance is on a different scale (different units) to the mean. If you didn't use the variance then ignore the previous sentence. The only options are the emulator standard deviation (square root of variance, i.e. mean +/- s.d.) or a 95% credible interval (i.e. an error bar whose lower bound is the 2.5th percentile and upper bound is the 97.5th percentile). A 90th credible interval (5th - 95th percentiles) is also fine.
(3) Figure 3 caption: After the word "uncertainty" please include brackets explaining what uncertainty you refer to, e.g. "The error bars indicate the uncertainty (95% C.I.) of the GP emulator output, ..." or the bit in brackets could be "...(mean +/- s.d.) ..." if that's what it refers to.
(4) Figure 4: in your responses to my comments you mentioned that you had changed "scalings" to "inputs" or "parameter" (I can't remember which you said without looking again) since not all of your inputs are scalings (e.g. ELEV). However the x axis labels for figure 4 you still use "scalings". Please can you fix this and also check for other possible general mentions of "scalings" (i.e. okay to refer to an input as a scaling factor if that's what it is but not as a general reference to any of the inputs).

Hide

ED: Publish subject to minor revisions (review by editor) (11 Oct 2018) by Paul Young

AR by Laura Revell on behalf of the Authors (14 Oct 2018) Author's response Manuscript

ED: Publish as is (02 Nov 2018) by Paul Young

Short summary

Global models such as those participating in the Chemistry-Climate Model Initiative (CCMI) consistently simulate biases in tropospheric ozone compared with observations. We performed an advanced statistical analysis with one of the CCMI models to understand the cause of the bias. We found that emissions of ozone precursor gases are the dominant driver of the bias, implying either that the emissions are too large, or that the way in which the model handles emissions needs to be improved.