the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 10 Feb 2026
Analysis of Antarctic ozone trends from 1979 to 2023
Abstract. Antarctic ozone has shown a sustained recovery since 2000, but levels were distinctly low during 2020–2023, potentially affecting estimates of ozone recovery and long-term trends. To assess the impact of recent low ozone on long-term variability, we analyze total column ozone (TCO) data from World Ozone and Ultraviolet Radiation Centre, multi-sensor reanalysis, and Total Ozone Mapping Spectrometer/Ozone Monitoring Instrument. Ozone fields from TOMCAT, a 3-D chemical transport model, are also used to gain better insight ozone changes. Multiple linear regression (MLR) is applied to estimate ozone trends over Antarctica from 1979 to 2023, incorporating proxies representing key chemical and dynamical processes such as the El Niño-Southern Oscillation and the Brewer-Dobson circulation (BDC).
Our analysis suggests that before 2000, all datasets show significant declines in annual TCO of about 2 and 6 Dobson Units per year (DU/yr) for September and October, respectively. For the 2001–2023 period, the magnitude of the October trend shifted to −1.5 DU/yr. The MLR effectively captures long-term ozone changes as well as unusual dynamical events such as the sudden stratospheric warmings in 2002 and 2019. As dynamical proxies show the largest influence, we use TOMCAT simulations to illustrate the impact of the BDC on the Antarctic ozone. Two sensitivity simulations further demonstrate that the strengthening (weakening) of the circulation leads to high (low) ozone values in spring. These findings suggest that after ozone-depleting substances were strictly controlled, dynamical processes have played an increasingly important role in controlling the ozone recovery patterns in Antarctica.
- Preprint
(1649 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 24 Mar 2026)
- RC1: 'Comment on egusphere-2026-560', Anonymous Referee #1, 11 Mar 2026 reply
-
RC2: 'Comment on egusphere-2026-560', Anonymous Referee #2, 16 Mar 2026
reply
Review of “Analysis of Antarctic ozone trends from 1979 to 2023”.
This paper investigates Antarctic total column ozone trends over 1979–2023 with a focus on recent unusually low ozone years and on the contributions of the BDC on Antarctic TCO variability. The paper is well written and mostly clear. Some of the figures do need to be reworked in some fashion to be able to be readable, especially by colorblind people. Justification for some of methodology needs to be included and I would like to see discussion of some of the results expanded to increase the novel aspects of the paper (see specific comments below). Overall, I think the paper will expand our understanding and knowledge of the Antarctic ozone hole recovery and will be a good fit for publication after the authors address the comments listed below.
Major comments
What is your justification for the time-period of your proxies used? Do your results show a difference if you use a 45 or 30 day lagged proxy instead of the Autumn-Spring accumulation for the eddy heat flux? Also, why is ENSO an instantaneous proxy? ENSO will affect tropospheric wave propagation which will likely take time to propagate down from the upper stratosphere. Does your ENSO proxy change the outcome if it is lagged?
Line 178, Table 3 and throughout. How much does this change if you just use 2000–2023 instead of 2001–2023? Other recent studies I have seen (including references cited in the paper) use 2000 as the starting point of ozone recovery (Steinbrech et al., 2017 Godin-Beekmann, etc.). What you have done is not necessarily wrong and it might not really change things, but some discussion on this somewhere in the paper is warranted. How important is end-point noise here?
In regards to modelling sections on the influence of the BDC. Have you looked into other potential conclusions here? Are there differences in the trends between the 2002 and 2006 years (EXP1 and EXP2)? If so what is causing them? Why is this analysis limited to only 10 years when the rest of the paper ends in 2023?
Please check the colorblind friendliness of your figures. There are a lot of separate lines, many of which are hard to distinguish (especially figure 5). I suggest reworking Figures 5, 8, and 9 to be more readable and to focus more on the conclusions you are trying to make. For example, using filled ranges with some key years overplotted to highlight specific conclusions could be one option.
Specific comments
Line 18-21. Please add in conclusions of the recent 2020-2023 years impact on ozone variability here too?
Line 53. Wang et al., (2025) may be an important reference missing here.
Line 64. The references on wildfire don’t talk about ozone depletion? There are other references that are likely more relevant here: Santee et al., (2022); Bernath et al., (2022); Solomon et at., (2023); Stone et al., (2025); Bruhl et al. (2025), etc.
Line 115-116. I suggest rewording this sentence as it is a little confusing as it stands: “TOMCAT is a three-dimensional chemical transport model (CTM) that simulates global data for stratospheric chemical elements and substances such as ozone based on consistent chemical equations”. By substances, do you mean aerosols, or VSLS, or just stratospheric trace gases in general? What do you mean by consistent chemical equations? You could just mention the source of the equation constants. I.e. JPL-xx.
Line 129. I understand what you mean here in a general sense but in the case of ozone the long-term trend isn’t really due to an unknown process?
Line 131. What does independent linear trend mean? From your equation 1 I assume you mean that you have two separate linear trends, but you may want to explain how this is different from PLT or why this technique is important to use here (or the benefits and flaws).
Line 134-135. I am confused by this sentence: “Not all aperiodic changes can be assumed to follow EESC, hence we employ ILT as the trend term of the MLR and ILT and all proxies match the entire period”. It sounds like you employ ILT as the trend term of ILT. I think you can split this into two sentences to avoid confusion.
Line 149. “To account for the effect of QBO phases and strength on ozone”. Do you mean: To account for the effect of the QBO phase on ozone variability?
Line 150-153. I think you should change to “stratospheric aerosol optical depth” like you have in Table 2 and change the acronym to SAOD to make sure you are clear that you are considering stratospheric aerosols only.
Line 159. “Are you really using AAO and ENSO as proxies for long term ozone change? You mention in the previous sentence that ENSO is related to early or late breakup of the polar vortex which isn’t really long term ozone change.
Line 173. I assume these are linear trends from the MLR and not standalone linear trends? I think you mention it is from MLR in Table 3 but might be best to say this here too.
Line 174. “In recent years”. I suggest giving the actual year range here.
Line 209. “Among them, the MLR of TOMCAT is the largest, explaining 91 % of the variance in the time series, indicating strong reproduction of observed long-term ozone variability.” This sentence is confusing. Among what? All panels of Figure 3 show TOMCAT only, and 91% of the variance is only for September. Also, I understand it is specified dynamics, but TOMCAT is a model, not observations. Please reword.
Line 249-250. I don’t think you can conclude that AOD dominates interannual variability here considering there is really only one event (Pinatubo) that is larger than any other proxies interannual variability.
Line 250. Is the 5.5% term referring to the maximum change in a given year or the explained variance?
Line 250-253: “The dominant role of BDC can be explained by its transport of ozone from the tropics to high latitudes, with ozone accumulation reaching a maximum at mid to high latitudes from May to September. During winter, ozone is transported from mid latitudes to the polar regions in spring, and the efficiency of this transport depends on the strength of BDC” These sentences seem to be explaining the same thing? Please refine.
Line 267. “Typical years”. I suggest removing typical as it may seem you are referring to normal years instead of more extreme BDC years.
Line 280-284. I don’t understand the reasoning here. I think this can be expanded a little. Where are the numbers 15% and 52% coming from? Also, the extremes in Figure 7 (blue and red lines) now show the linear trends predominantly dominated by ODS recovery similar to what was done by Solomon et al., (2016). Have you looked into other results here? Are there differences in the trends between the 2002 and 2006 years? If so what is causing them?
Line 325. Maybe self-evident, but I suggest mentioning that this change in trend sign is due to large ozone holes later in the period (after 2019).
Line 326. “explaining about 67-91 %”. Is the large range due different months? 65% in September and 91% in October?
Figure 1 title. Correlation between proxy -> correlation between proxies
Figure 2. I think you can keep the same x-axis limits for panel d to make it easier to compare with the other datasets. Also, y-axis limits should be the same here.
Figure 4. Is there any particular reason for the dates shown on the x-axis? I like the vertical dashed line separating the months but the 20-day tick mark is unusual?
Figure 3. What does OLS mean in the legend? Ordinary Least Squares maybe? I can’t find this acronym throughout the text and am confused why ordinary least squares would be used here (it that is indeed the case).
Figure 5. Please restructure this figure. It is very hard to distinguish between lines here. It will be virtually impossible for a colorblind person. If there is a supplement, this figure could be a good fit considering that most information is captured in Figure 6.
Figure 7-9. It looks like these figures have stretched aspect ratios? I assume it will be fixed later. Please check.
Figures 8 and 9 are going are also going to be hard for colorblind readers. I suggest reworking. Maybe using a filled range with only extremes and notable years overplotted. (I understand that it is not an easy ask).
Figure 9. I am confused why this analysis only goes to 2009 in contrast to the rest of the paper which extends to 2023?
Technical corrections
Line 109. Total ozone column – TCO for consistency.
Line 133. “changes in the ODS” -> “changes in ODS”
Line 154. “use Bremen” to “use the Bremen”
Line 177. “2001- 2023”. Remove space and use endash.
Line 207. “Explaining” -> “Explain”
Line 254. “of BDC” -> “of the BDC”
Line 255 “peak rate” -> “the peak rate”
Table 1. EAR5 -> ERA5 (I assume).
Line 260-261. “To investigate the role of BDC on Antarctic ozone” suggest changing to “To investigate this further”.
Line 295. “The EXP1” -> “EXP1”
References
Wang, P., Solomon, S., Santer, B. D., Kinnison, D. E., Fu, Q., Stone, K. A., Zhang, J., Manney, G. L., and Millán, L. F.: Fingerprinting the recovery of Antarctic ozone, Nature, 639, 646–651, https://doi.org/10.1038/s41586-025-08640-9, 2025.
Santee, M. L., Lambert, A., Manney, G. L., Livesey, N. J., Froidevaux, L., Neu, J. L., Schwartz, M. J., Millán, L. F., Werner, F., Read, W. G., Park, M., Fuller, R. A., and Ward, B. M.: Prolonged and Pervasive Perturbations in the Composition of the Southern Hemisphere Midlatitude Lower Stratosphere From the Australian New Year’s Fires, Geophysical Research Letters, 49, 1–11, https://doi.org/10.1029/2021gl096270, 2022.
Bernath, P., Boone, C., and Crouse, J.: Wildfire smoke destroys stratospheric ozone, Science, 375, 1292–1295, https://doi.org/10.1126/science.abm5611, 2022.
Solomon, S., Stone, K., Yu, P., Murphy, D. M., Kinnison, D., Ravishankara, A. R., and Wang, P.: Chlorine activation and enhanced ozone depletion induced by wildfire aerosol, Nature, 615, 259–264, https://doi.org/10.1038/s41586-022-05683-0, 2023.
Stone, K., Solomon, S., Yu, P., Murphy, D. M., Kinnison, D., and Guan, J.: Two-years of stratospheric chemistry perturbations from the 2019/2020 Australian wildfire smoke, Atmos. Chem. Phys., 25, 7683–7697, https://doi.org/10.5194/acp-25-7683-2025, 2025.
Brühl, C., Kohl, M., Lelieveld, J., Rieger, L., and Santee, M.: Radiative forcing and stratospheric ozone changes due to major forest fires and recent volcanic eruptions including Hunga Tonga, , https://doi.org/10.5194/egusphere-egu25-3642, 2025.
Solomon, S., Ivy, D. J., Kinnison, D., Mills, M. J., Neely, R. R., and Schmidt, A.: Emergence of healing in the Antarctic ozone layer, Science, 310, 307–310, https://doi.org/10.1126/science.aae0061, 2016.
Citation: https://doi.org/10.5194/egusphere-2026-560-RC2 -
CC1: 'Comment on egusphere-2026-560', Hannah Kessenich, 17 Mar 2026
reply
Kessenich et al., 2023 is cited in the context of the large and long-lasting ozone holes during 2020-2023. However, in the context of the work presented in the manuscript, this publication should be cited for its finding of negative daily ozone trends in October.
Figure 4 in this manuscript, for example, appears to replicate Figure 3 from Kessenich et al., but now with TOMCAT data rather than MLS/Aura observations. The TOMCAT results appear to closely agree with the MLS results presented in Kessenich et al, but this is not mentioned.
Good scientific practise requires that appropriate credit should be given when an analysis is adapted or extended. As such, please add appropriate citation to the work of Kessenich et al. 2023.
Citation: https://doi.org/10.5194/egusphere-2026-560-CC1
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 210 | 146 | 23 | 379 | 44 | 51 |
- HTML: 210
- PDF: 146
- XML: 23
- Total: 379
- BibTeX: 44
- EndNote: 51
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The manuscript “Analysis of Antarctic ozone trends from 1979 to 2023” by He et al. provides an
analysis of the long-term evolution of total column ozone over Antarctica from 1979 through 2023
based on four data sets: WOUDC ground-based measurements, the Multisensor Reanalysis (MSR-2),
the Total Ozone Mapping Spectrometer/Ozone Monitoring Instrument (TOMS/OMI) record, and
Chemical Transport Model (CTM) simulations from TOMCAT. The authors apply a standard multiple
linear regression (MLR) approach taking into account various atmospheric key processes which affect
ozone variability. They focus on the investigation of the role of the Brewer-Dobson Circulation (BDC)
on ozone changes, and they perform two sensitivity experiments with TOMCAT also focused on the
role of the BDC. The latter is limited to the period 2000-2009. On top of that, the authors investigate
the divergence of the September and October trends.
The topic of the manuscript fits into the scope of ACP. However, I think that the novelty of the
investigation is somewhat limited, but I would recommend publication after revision.
Major comments
In the abstract, you mention the distinctly low ozone values during 2020-2023 and the intention to
assess the impact on long-term trends. Unfortunately, this assessment is limited to a short paragraph
in Sec. 4. I recommend to elaborate on this in more detail, in particular with respect to the possible
reasons for the reduced values, e.g. the exceptional atmospheric conditions during these years.
Related to the previous point: Why did you limit the model sensitivity experiments to the period
2000-2009? The correlation between BDC and polar ozone for 1995-2010 was already investigated in
detail in Weber et al. (2011). I think, analysing the behavior in the last years (2020-2023) would be
interesting. Is there any possibility to extend the model analysis?
Minor comments
p. 3, l. 89:
consistancy -> consistency
p. 4, Sec. 2.1:
How many ground-based stations are located in the latitude band 60°S-90°S?
p. 4, l. 100:
Please use “MSR-2” consistently throughout the manuscript.
p. 4, ll. 102-103:
MSR-2 also includes SBUV/NOAA-17, -18, -19.
p. 4, l. 107:
For MSR-2, use “van der A et al., 2015”.
van der A, R. J., Allaart, M. A. F., and Eskes, H. J.: Extended and refined multi sensor reanalysis of total
ozone for the period 1970–2012, Atmos. Meas. Tech., 8, 3021–3035, https://doi.org/10.5194/amt-8
3021-2015, 2015.
p. 4, Sec. 2.3 TOMS/OMI:
I suggest to delete the first sentence since EP TOMS was decommissioned in 2007. The datasets for
TOMS and OMI are two separate data records with different spatial resolutions. Please describe, how
you merge them for the analysis. Did you compare them during their overlap period? Did you apply
any adjustment to one of the records in order to avoid artificial jumps? Please provide some more
details here.
p. 5, Table 1:
The link https://woudc.org/archive/Projects-Campaigns/ZonalMeans provides only zonal means for
the period until 2021 (file gb_1964-2021_za.txt). What is the source for the extended (until 2023)
record?
3rd column (spatio-temporal resolution): Please (i) provide the resolution in degree [°] for MSR-2,
TOMS/OMI, and TOMCAT, (ii) remove “lat*lon=….” in the second and third row, and (iii) explain T42
L32 for TOMCAT.
4th column, last row: “EAR 5” -> “ERA5”
p. 5, 133-134:
“These trend terms represent the only non-periodic terms of MLR…” -> Not sure, if this statement is
correct. For example, AOD is non-periodic as well. Please explain what you mean here.
p. 6, Eq. (1) and following text:
Please explain all terms, e.g. QBO_10(t), QBO_30(t), S(t), E(t), ….
Does “t” represent the month or the year?
p. 6, l. 160:
“…while other proxies use the monthly mean time series” -> In line 143, you indicate that “t”
represents the year and not the month. Please clarify. Maybe “t” represents the month (see my
previous comment).
p. 6, Table 2:
Please check URL for QBO indices and BDC.
p. 8, l. 172:
“Antarctic ozone recovery exhibits strong seasonal dependence, particularly the contrasting behavior
in September and October” -> Please provide some more explanation for this statement and a
reference.
p. 8, l. 177:
“September shows signs of recovery” -> According to table 3, all trends are very close to zero and
statistically not significant. I suggest to mention this here. The same holds for October; the negative
trends are statistically not significant.
p. 9, ll. 190-194:
September trends are positive for both periods (2001-2019 and 2001-2023), but for 2001-2023, the
trends are very close to zero and reduced by a similar value as the October trends. Dynamical
processes do not only affect October values (see also your Fig. 7).
p. 11, l. 206:
“September shows a weak positive trend of 0.1” -> I suggest to rephrase. This is rather “close to zero”
than positive. And it has a large 2-sigma uncertainty (1.8DU/year).
p. 11, l. 207:
“can explaining” -> I suggest to rephrase, e.g.: “The independent variables in the MLR can explain
about 85%...”
p. 12, ll. 208-209:
Can this be also related to the possibility that models do not entirely capture the complete variability
of the atmosphere?
Line 209: Replace “observed” with “simulated”.
p. 12, l. 220:
“However, positive trends dominate the middle and lower stratosphere” -> replace “middle” by
“upper”; positive trends are found around 2-3 hPa.
p. 14, l. 238:
“, with BDC being the main driver of long-term ozone changes” -> I would suggest to rephrase: “, with
BDC being the main driver of interannual ozone variation and an important contributor to long-term
ozone changes.”
p. 14, Fig. 6:
Do I understand correctly, that these peak contributions shown in Fig. 6 were obtained from the
curves in Figure 5? If so, I do not see peak contributions of the BDC of 80-100DU (Fig. 6) in the curves
of Fig. 5 (b) and (c).
Please explain, what you mean with “rate of ozone change [in percent]”?
p. 17, Figure 7:
Title: the period shown here is 2000-2009.
p. 18, Fig. 8(b) and p. 19, Fig. 9(b):
The year 2000 is quite different in EXP1; lower ozone values compared to the subsequent years, but
EHF (Fig. 9b) is also quite low in October (green curve). Do you have an explanation?
p. 19, Figure 9:
(i) Title of top panel: “Tempetature” -> “Temperature”
(ii) I would suggest to highlight the exceptional years 2002 and 2006 in both panels a bit more for
better visibility.
p. 19, l.306:
“(Fig. 8 and Fig. 9a)” -> Do you mean Fig. 7 and Fig. 8a?
p. 20, ll. 331-332:
I suggest to rephrase this sentence. Please explain the impact of the BDC. The start a new sentence
for volcanic aerosols.
p. 23, ll. 395-397:
Reference refers to article in ACPD. Replace with reference to final revised version in ACP.
Dhomse, S. S., Kinnison, D., Chipperfield, M. P., Salawitch, R. J., Cionni, I., Hegglin, M. I., Abraham, N.
L., Akiyoshi, H., Archibald, A. T., Bednarz, E. M., Bekki, S., Braesicke, P., Butchart, N., Dameris, M.,
Deushi, M., Frith, S., Hardiman, S. C., Hassler, B., Horowitz, L. W., Hu, R.-M., Jöckel, P., Josse, B., Kirner,
O., Kremser, S., Langematz, U., Lewis, J., Marchand, M., Lin, M., Mancini, E., Marécal, V., Michou, M.,
Morgenstern, O., O'Connor, F. M., Oman, L., Pitari, G., Plummer, D. A., Pyle, J. A., Revell, L. E., Rozanov,
E., Schofield, R., Stenke, A., Stone, K., Sudo, K., Tilmes, S., Visioni, D., Yamashita, Y., and Zeng, G.:
Estimates of ozone return dates from Chemistry-Climate Model Initiative simulations, Atmos. Chem.
Phys., 18, 8409–8438, https://doi.org/10.5194/acp-18-8409-2018, 2018.
p. 23-24, ll. 427-430:
Something seems to be wrong with initials, please check.
p. 26, ll. 504-505:
Year is missing.