Review of Preprint egusphere-2023-565

Weakening of springtime Arctic ozone depletion with climate change 

by Marina Friedel, Gabriel Chiodo, Timofei Sukhodolov, James Keeble, Thomas Peter, Svenja Seeber, Andrea Stenke,Hideharu Akiyoshi, Eugene Rozanov, David Plummer, Patrick Jöckel, Guang Zeng, Olaf Morgenstern, and Béatrice Josse, https://doi.org/10.5194/egusphere-2023-565

Overview:

This paper addresses the issue whether the combination of decreasing halogenated ozone-depleting substances and increasing greenhouse gases will lead to changes in the frequency of major Arctic ozone minima by the end of the 21^st century.  Complicating the matter is that there are large differences in CCMI (1&2) models, not all of which are able to simulate the magnitude of Artic ozone minima in the current climate.  They find that models that overpredict ozone minima in the present climate (those with a cold pole bias) have a decrease in the number of ozone minima under a future GHG scenario.  Those that are warm biased have small sensitivity to changes in both GHGs and ODS concentrations.  Overall, they find the Arctic ozone minima will lessen with increases in GHGs, largely due to decreases in ODSs.  The stratospheric cooling caused by increases in GHGs that could potentially increase ozone minima is weakened by opposing radiative and dynamical mechanisms.

It is found that models give different answers because they have different sensitivities to GHG and ODS forcings, in particular with respect to lower stratospheric transport and dynamics, and thus have a large inter-model spread in temp and ozone trends.  To come to their conclusions, given all the different model results, the authors compared the ozone evolution across different CCMs and GHG emission scenarios, while identifying reasons for model discrepancies. They then linked the model spread in future ozone trends to differences in model climatologies,  and compared with observations to identify the likely evolution of future Arctic ozone minima. 

---------------------------------------------------------------------------------------------------------------------------------------------------------------

1) Page 2, line 33;  this references WMO 2018…the authors may want to check whether the statement is still supported by WMO 2022.

2) page 3, line 65, add a comma after “two CCMs” and after “version 4”

3) question, why do you normalize to the period mean ozone?  From a surface UV point of view, it’s the absolute ozone value rather than a deviation from average that matters.   So, it seems one should identify the ozone minima based on an absolute DU threshold rather than a deviation from a period climatology.  At a minimum, in Figure 1, I’d include a notation of what the mean used is.  And, it seems that the horizontal scale for the multimodel means (b&d) is different from the individual models, and for ease in looking at the figure, it should be the same.

4) This is related to the normalization: on page 7 line 165, you note looking at ozone minima in a 25-year running window.  What do you use for the base ozone for identifying the extreme ozone minima events?  Is it still one value for the start of the time serios, and a second for the end?  Are there less events at the end of the time series? (shown in Figure 2)

5) Page 8, line 210:should say a reduction in the strength of the BDC.

6) Page 10, line 229: Rather than saying agree better with reanalysis (which is rather generic) I recommend saying  agree better with MERRA 2 ozone if that’s what you compared with.  Would be even better to compare ozone with satellite measurements (such as MLS).

7) Page 13, line 283 says “the contribution of individual processes to temperature trends might vary across models”  It seems that, with the model output you have, you can show this.  Take 2 different models with different trends and analyze the output to see what’s driving the temperatures trends.

8) Page 13, conclusions…Does von der Gathen 2021 really question the reliability of simulated ozone?  They really don’t even use simulated ozone, instead they use assorted proxies. And after rereading Morgenstern et al., I don’t think it supports questions the reliability of simulated ozone.

9) This recent paper is relevant to this study:  http://www.columbia.edu/~lmp/paps/polvani+etal-NATURECOMM-2023.pdf  (Polvani et al., No evidence of worsening Arctic springtime ozone losses over the 21st century )  One of their conclusions is “When all the relevant process are included, as they are in the state-of-the-art comprehensive chemistry-climate models, there is no evidence that future ozone levels will decrease in the coming decades, including over the Arctic in springtime, as we now explicitly show.”  This seems to be very similar to the conclusions of this study. This should be refereced around page 15, line 316.

Friedel et al. provide a comprehensive study on the future evolution of ozone minima over the Arctic polar cap. The authors present a thorough analysis of a series of transient and timeslice simulations with 2 CCMs (WACCM and SOCOL) as well as the cohort of simulations available from CCMI1 and CCMI-2022 models. In their study the authors: i) address individual model weaknesses (warm and cold biases) related to the realization of ozone minima; ii) explore the related spread in modelled springtime ozone anomalies at present and for different future climate scenarios; iii) quantify the magnitude of ozone anomalies for early and late 21^st century and temporal changes in these anomalies; iv) illustrate how the amount of ClOx available in CCMs drives ozone minima at present and also determines their future trends; and v) detail how inter-model spread can be used to constrain ozone minima projections.

The manuscript is timely and well prepared. The study sheds new light on the long-standing question regarding the future evolution of Arctic low ozone extremes and emphasizes the central role of declining ODS abundances for future Arctic ozone, both mean and extreme, across potential climate pathways.

I recommend accepting this manuscript for publication after addressing the comments provided below.

General comments:

Section 3.1, L163-170: The authors examine ozone minima present and future, and the model spread in ozone minima across CCMs. While the text and Fig. 2 illustrate well the overall change in ozone anomalies I would suggest also including a short text passage or supplemental table detailing the timing of the most pronounced ozone minima per CCM and scenario (on decadal basis).

Section 3.3, L210-228: I would assume the picture would not change strongly but how would Fig. 3 look like if you restrict to ClOx and temperature in spring (March-April)?

Section 3.3, L261-264: I would call the correlation between stratospheric temperature trends and ozone minima of 0.59 moderate not weak. However, I agree with the dominance of some models for overall R. Thus, I would recommend specifying how R changes if the two most extreme (positive and negative) models are removed.

Section A2: I agree with the authors to apply model weighting, however given the substantial difference in weight for different models across scenarios I would suggest adding a second panel to Fig. A5 showing also the evolution of the unweighted multi-model mean of ozone minima strength.

Technical comments:

L 155: on a multimodel mean --> on the multimodel mean

L166: Development --> evolution | (also in caption of Fig. 2)

L210: reduction of the BDC --> weakening of the BDC

L273: Fig.5 --> Fig. 5

Figure 1, caption: normalized by mean ozone --> normalized by mean partial column ozone

Fig. 2, caption: normalized by ozone climatology --> normalized bz the ozone climatology

Fig. A7: is the vertical whisker for CMAM ClOx missing?