the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 May 2026
Measurement report: Significant ozone loss during winter 2020 measured from ground based microwave radiometer
Abstract. Ground-based microwave observations from MIRA2 situated in Kiruna, Sweden, were used to investigate Arctic stratospheric ozone during the winter 2019/2020. A comparison of O3 retrievals with coincident measurements of Aura MLS between 1 October 2019 and 30 April 2020 shows good agreement across the investigated pressure levels (74, 56, 46, and 10 hPa). Remaining differences are well within the retrieval uncertainty of MIRA2. This demonstrates the capability of MIRA2 to provide robust ozone measurements for studies of stratospheric variability.
A tracer-based approach was applied to derive cumulative chemical ozone loss on isentropic surfaces. At the 475 K level, ozone depletion increased from late winter into early spring, reaching a maximum loss of 2.19±0.90 ppmv in early April 2020. The magnitude and timing of the loss are consistent with the exceptional Arctic ozone depletion stated by model simulations and satellite-based estimates during the winter 2019/2020 and agree well. Despite limited temporal sampling, the tracer-based method enables a consistent estimate of seasonal ozone loss from ground-based observations. The results highlight the ability of ground-based microwave radiometers to quantify chemical ozone depletion and its temporal evolution.
This study also demonstrates the value of ground-based measurements as an independent and complementary component of the global observing system, providing continuity and reducing reliance on satellite data for monitoring stratospheric ozone.
- Preprint
(7232 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-2216', Anonymous Referee #1, 17 May 2026
-
RC2: 'Comment on egusphere-2026-2216', Chris Boone, 28 May 2026
This paper shows interesting results from ground-based microwave (MIRA2) measurements of O3, using atmospheric tracer information (in the form of N2O) from MLS to derive catalytic ozone losses within the polar vortex. Overall, the study is clearly described and is worthy of publication. I had a relatively minor issue with some of the discussion regarding things happening inside the polar vortex while MIRA2 was measuring outside the vortex, where the discussion seemed to imply that the things happening inside the vortex were shown in the figures, when they were not, as I will address later. I would also have liked a discussion of how the ozone loss analysis approach might work once MLS reaches end of life and its N2O measurements are no longer available to use as a tracer.
Perhaps there could be a few more words regarding details of the data set. What is the sampling time? Do measurements require extensive averaging? Are measurements taken continuously, day and night, or are only a few measurements collected per day? Continuous operation would improve the chances of a satellite mission finding a good coincidence, so that might be useful information. Observations in the paper are distilled down to focus solely on coincidences with MLS. I assume the data set is more extensive than that. I could browse through the data to see what it held, but it would be helpful to provide a sense of that in the paper describing the measurement set. If the data consisted of relatively short measurement times and a variety of elevation angles, one might consider analyzing multiple measurements simultaneously to potentially improve the altitude resolution. Conversely, if measurement times are long and you don’t have multiple observations over a relatively short period of time at a variety of elevation angles, that might not be a viable option. From the paper itself, without browsing the data, I can’t tell what the sampling or extent of the data set might be.
You might consider saying “Arctic winter” in the title rather than just “winter.”
>Line 77: In the above mentioned frequency range, observation of HNO3, N2O, ClO and O3 is possible. However, due to instrumental limitations MIRA2 has been operated continuously in ozone observation mode only.
As discussed several times in the article, MLS is not expected to continue producing data for much longer. The ozone loss analysis is completely reliant on an outside source of coincident measurements of O3 and at least one atmospheric tracer (like N2O). Without that information, it is not possible to separate chemical losses from transport effects. However, the above statement seems to imply that the microwave radiometer itself could also measure N2O, which would remove the need for an outside source of measurements from MLS. It wasn’t clear to me what the instrumental limitations were. Is it possible to have a self-contained ground-based instrument that could derive ozone loss using its own set of measurements (and have ClO on top of that, to infer chlorine processing)? I wondered why this capability was mentioned but never explored.
>Line 261: Measurements within the polar vortex are indicated by red crosses, and white markers denote dates when MIRA2 sampled vortex air masses.
In Figure 5, there are red squares (not crosses) and no white markers that I can see.
>Line 276: During February and March, these periods also show enhanced ClO volume mixing ratios
In Figure 5, there are no enhancements in ClO seen in March, but judging from the red squares the measurements in March are outside the polar vortex. The discussion appears to strongly imply that Figure 5 shows evidence of chlorine activation in March, but it does not because it reports MLS ClO coincident with MIRA2, which is measuring outside the vortex, where ClO levels are close to zero.
>Line 283: The air masses sampled by MIRA2 have undergone substantial depletion during the period of peak ClO activation in mid-March 2020
My understanding is that one typically refers to chlorine activation, not ClO activation. Again, nothing presented in this paper substantiates that ClO inside the vortex peaked in March. It would be advisable to add a figure showing ClO measured by MLS within the polar vortex (not outside the polar vortex in March, like Figure 5) or provide a reference to a paper where that information has explicitly been shown.
The idea is that the Arctic polar vortex was (as usual) not stationary, as mentioned in Section 3.3. The ground-based site was sometimes inside and sometimes outside the vortex because the vortex was moving. During March 2020, the MIRA2 site was outside the vortex, at which time significant chlorine processing destroyed ozone inside the vortex, but that was “off-stage” to MIRA2 with its measurements being outside the region of ozone destruction. In April, the vortex moved such that the MIRA2 site was once again back inside, at which point the MIRA2 measurements saw the aftermath of the chlorine processing, reduced ozone levels. This is addressed to some extent in the text, but there is also discussion of things happening off-stage as though the figures are showing information on chlorine processing in March, which they are not. It would be good to have a clearer distinction in the text as to what relates to the figures and data provided in the paper, and what relates to “outside” knowledge (like ClO peaking in mid-March).
>Line 285: By early April, ClO has largely been deactivated
Again, my understanding is that it would be phrased chlorine deactivation, not ClO deactivation.
Minor (mostly wording) issues
-----------------------------
Should the Abstract perhaps be a single paragraph?
>Line 20: During the second half of the 20th century holes in the ozone layer began to form due to Ozone Depleting Substance (ODS)
The wording here makes me picture a Swiss-cheese type atmosphere with persistent holes. The ozone hole forms in the Antarctic during Southern Hemisphere winter/spring but goes away when the polar vortex breaks up. There was also overall thinning of the ozone layer globally. I do not get that picture from this statement.
>Line 22: The absolute majority of the ODS gasses does not have a natural origin
The word “absolute” seems unnecessary
>Line 23: This prompted the Montreal Protocol, banning industries and other actors to produce ODS
The Montreal Protocol did not ban production; it gradually phased it out.
>Line 58: denitrification through sedimentation of HNO3 reduces reactive nitrogen
The wording seemed odd to me, in that it is not gaseous HNO3 that is sedimenting. I would instead say something like sedimentation of HNO3-containing PSC particles. That is probably just personal preference.
>Line 98: observations at a given site may both sample air inside the vortex, within the vortex edge region, or outside the vortex altogether
I would remove the word “both.” It implies two options, but three are given.
Citation: https://doi.org/10.5194/egusphere-2026-2216-RC2 -
AC1: 'AC response to RC1', Richard Johansson, 08 Jul 2026
We thank the reviewer for taking the time to review our manuscript. The comments were constructive and very relevant. We will begin by addressing the reviewers major comments:
Information on passive ozone: Some very valid points were raised considering our section describing how passive ozone was obtained. We agree that clarification regarding how we obtained the passive ozone might be necessary.
1. We rewrote much of section 4.6 to clarify how the reference function between N2O and O3 was estimated. We now include the function we fitted to the samples of N2O and O3, how that was done and also elaborated on why we choose December 9 2019 for this.
2. We now elaborate on appropriate timescales for this reference function in Section 4.6.
3. We included a figure for the samples used for the reference function, and the resulting fit (Figure 3).
4. As suggested we plotted the passive ozone in Figure 6, and also added estimated pressure levels for 475 K.
Claims that we could independently quantify ozone loss only from MIRA2 data: As the reviewer pointed out, we depend on N2O measurements, which we obtained from MLS in thisstudy. We therefore can not independently estimate chemical O3 loss solely from the measurements we obtain from MIRA2. We have now clarified this at relevant places in the manuscript;
1. Last part in the abstract, where we now explicitly explain that we actually depend on MLS N2O data to estimate chemical O3 loss. We also elaborate on what options we have in the future, when MLS no longer is operational.
2. In the last paragraph of the introduction we also had this claim, we now have corrected this, where we now instead explicitly say that we use N2O from MLS to estimate the chemical O3 Loss
3. We also changed the last paragraph of Section 5.2, where we now also clarify that we do depend on N2O from MLS in this study to do the estimates of chemical O3 loss.
Instrument temperatures: Regarding the comment on the instrument temperatures we would like to explain the following;
1. We cool down the mixer and the first amplifier to roughly 70 K with a closed-cycle two-stage He refrigerator. We have clarified this in the manuscript now.
2. We state that MIRA2 has a system noise temperature Tsys = 1250 K. This is the brightness temperature that the receiver ’sees’ due to the noise of the electronic components, therefore called system noise. When it comes to the system sensitivity, i.e. the smallest signal ∆Tmin (from the stratosphere) that can be resolved, the system noise as a fixed property is counterbalanced by two other factors, the effective bandwidth of the spectrometer channels and the integration time. The radiometer equation, ∆Tmin = Tsys/√Bt , describes the relationship between these three magnitudes. With Tsys = 1250 K and an effective channel bandwidth B of 180 kHz an integration time t of roughly 217 s is needed to achieve a ∆Tmin of 0.2 K, which is more than is needed for resolving a stratospheric ozone emission line. By either choosing a spectrometer with less spectral resolution or by increasing the integration time, ∆Tmin can be improved.
~
The above is our response to the reviewers major comments. The response below refers to the reviewers minor comments and the bold Line/Figure refer to the line and figure in the reviewed manuscript:
Lines 6, 256: We added the approximate pressure levels for 475 K.
Line 26, 31: We added the suggested reference and changed our wording considering the recovery.
Line 41-42: As suggested we limit the introduction part to the focus of our study. We want to emphasize, however, that with the end of MLS operations the need for ground-based measurements becomes even more important.
Line 58: As commented, the loss cycle starts with Cl, and is now fixed. However, both Cl and ClO radicals are important in the ozone depletion reactions.
Line 59: As suggested, we now use active forms in this sentence instead.
Line 76: We changed the wording here to use "retrieved" instead, also slightly adjusted the ranges and added a reference for this. As you say, a narrower bandwidth would not give us more information, rather the opposite. We refer to Ryan et al., 2016.
Line 84: We added a reference (Jiang et al., 2007) here where MLS has been thoroughly validated against in situ measurements and other satellite measurements.
Line 117: We slightly changed the wording here.
Figure 1: We now indicate the period for which the coincident measurements were obtained.
Line 154: e added the missing word "loss" after “O3”
Line 201: As suggested we added "vertical" before transport, and this is an important distinction.
Line 204: Yes, meteorological fields in the DMP dataset come from MERRA-2, and we now explicitly state this. The reference for this is through the DMP dataset reference.
Figure 5: We added passive ozone in subplot b, which now is in Figure 6
Line 279, 280: We followed the suggestion to drop this sentence.
Figure 6: This comment refers to Fig 7 now. We added maximum loss found by Manney and Wohltman in this figure as stars. We also added approximate pressure level for 475 K in the figure title. While be agree that it would illustrative to plot ozone depletion over time found in previous studies, the data to do this is not easy to obtain especially with respect to our tracer-tracer approach.
Line 289: We rephrased this sentence according to the suggestion.Citation: https://doi.org/10.5194/egusphere-2026-2216-AC1 -
AC2: 'AC response to RC2', Richard Johansson, 08 Jul 2026
We thank the reviewer for taking the time to carefully review our manuscript. We greatly appreciate the constructive and insightful comments, which have helped us improve the quality and clarity of the manuscript. We will first address the reviewers major comments:
MIRA2 measurements of other species than O3: There are signatures of N2O in the frequency range of MIRA2. However, these faint signatures are hard to resolve given the overall performance of MIRA2 and have been clarified in the manuscript. There is no existing ground-based microwave spectrometer (at IRF) which is capable of measuring N2O of which vertical profiles can be retrieved, which would be necessary for our tracer approach to determine the chemically induced ozone depletion. Other EOS observations would be preferable.
References to the markers in Figure 5: The sentence mentioning red crosses and white markers referred to a previous version of Figure 5 and was inadvertently overlooked prior to submission. We have removed this sentence in the revised manuscript, as the figure caption already provides a clear explanation of the red squares shown in the current version of the figure, which now is Figure 6.
Discussion of Figure 5: We have revised much of the section where we discussed what was shown in Figure 5 (now Figure 6). What we have changed is:
1. We have slightly reorganized the structure to improve the overall flow in section 5.2
2. We now draw a clear distinction between what we directly interpret from the data we showin Figure 6 (previously Figure 5), and what is derived from earlier studies, primarily from
Wohltmann and Manney, with the corresponding references included.
3. We have revised the text in light of our decision to refer to it as “ClO activation” and haveadopted the suggestion to instead use the term “chlorine activation/deactivation”.
Discussion regarding details of the data set: We have now in the revised manuscript given more details of the data set used from MIRA2. But to answer some of your questions:
1. “What are the sampling time of the measurement?”: MIRA2 measurements consist of roughly 1-4 hours of observation performed continuously.
2. “Do the measurements require extensive averaging?”: No, unless the reviewer sees a three-hour integration time of one sample as extensive. However, for some coincidences we have up to three shorter MIRA2 measurements which we have averaged ensuring that the measurement time matches the temporal coincidence criteria (±2 hours around the MLS measurement)
3. “Are measurements taken continuously?”: Yes, MIRA2 measurements are obtained continuously in samples of one to four hours depending on variations in the tropospheric conditions. At the beginning of each new sample the elevation angle is chosen according to the actual tropospheric conditions.
~
The above refer to the reviewers major comments. We will address minor comments below, and
The bold text refer to the Line/Figure/Section in the reviewed manuscript:
Should the Abstract perhaps be a single paragraph?: We have now adjusted the abstract and put all of it into one paragraph, since this seems to be the standard in ACP.
Line 20: We have substantially reworked this paragraph to provide a much clearer account of what is happening, beginning with Farman’s first observations and tracing developments through to the signing of the Montreal Protocol.
Line 22: Based on the revisions we made in response to the comment on line 20, this comment has now also been resolved.
Line 23: We have changed the wording here in accordance to the suggestion.
Line 58: We have rephrased this to reflect that this occur on PSC surfaces
Line 98: As pointed out, the way we wrote this would suggest two options. We therefore dropped the word "both" since there are three options.Citation: https://doi.org/10.5194/egusphere-2026-2216-AC2
Data sets
MIRA2 measurement and retrieval data from 20191001-20200501 R. Johansson, U. Raffalski, and J. Groß https://zenodo.org/records/19608988
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 300 | 48 | 11 | 359 | 27 | 24 |
- HTML: 300
- PDF: 48
- XML: 11
- Total: 359
- BibTeX: 27
- EndNote: 24
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The paper describes stratospheric ozone profile data obtained in winter 2019/2020 by a microwave radiometer at Kiruna, northern Sweden. The ground-based microwave data are compared with and supplemented by data from the satellite-based microwave limb sounder instrument. Ozone profiles from both instruments compare well over the pressure / altitude range from 70 hPa to 1 hpa. Using a tracer relationship between O3 and N2O from the MLS data, the authors then compare observed O3 with passive O3 from the tracer relation. This provides estimates of the large chemical ozone loss that occured over the Arctic in the unusually cold stratopsheric winter of 2019/2020. The derived ozone loss is consistent with previous studies.
While there are no major new findings this is still a well written paper that fits into the scope of ACP. I recommend publication after a few generally minor revisions.
My most major comment is about lack of information on how the passive ozone time series was derived. I think it is necessary to provide a plot (scatter plot?) of the underlying ozone vs. N2O data from MLS, before or after Fig. 5. More statements about the time period for the used relation, and how this would change over time is also needed. In Fig. 5 it would be good to also see the passive ozone plotted.
Also, in various places in the manuscript, the authors claim that MIRA data can independently quantify chemical ozone depletion and its temporal evolution (e.g. lines 11, 37, 298/299, 316/317). I question this claim, because the approach relies heavily on the N2O to ozone relation as well as on the N2O profiles provided by MLS. Without MLS or comparable data, I have a hard time seeing how the authors could derive ozone loss from the MIRA data alone. This claim either needs information about available ground-based data that support the approach, or the claim needs to me modified.
In Fig. 6, I am missing additional lines / more information for the ozone depletion over time found by Manney et al. 2020, and Wohltmann et al. 2021.
Lines 6, 256: maybe add 60 to 50 hPa somewhere after 475 K, to give additional information about the altitude range.
Line 26, 31: I suggest to add the 2022 WMO/UNEP Scientific Assessment of Ozone Depletion as a reference. In line 31, it is not clear whether lower stratospheric ozone is actually increasing, although this is expected. Maybe better to say: Although the decline of stratospheric ozone has been stopped, ...
Line 41-42: While I strongly support ground-based measurements, I don't think your work has done a lot to "highlight the need to ...". I suggest to formulate differently, and stick with what you have actually shown.
Line 58: active chlorine is Cl, not ClO, and the loss cycles start with Cl
Line 59: I would use the active forms: limits and prolongs
Line 72: I don't understand these temperatures. You cool the mixer diode to 70 K? With what? Liquid nitrogen that has 77 K? What is the meaning of 1250 K? I would assume that with 1250 K noise you would not be able to measure anything in the stratosphere, which has about 250 K. Please explain.
Line 76: Instead of "achieved", I would say "retrieved" If you had narrower bandwidth, you would probably not be able to derive much more atmospheric information? Also, this sentence needs a reference.
Line 84: Again, I think a reference is necessary here.
Line 117: instead of "and their associated uncertainties" I would say "on the basis of their associated uncertainties"
Fig. 1: Please indicate which time period is considered here. One day? One month?
Line 154: loss missing after O3
Line 201: I think you need to add "vertical" before transport. Horizontal transport on isentropic surface can and will change ozone. To a large degree you exclude that by only using high PV / vortex data, but it could happen.
Line 204: Which meteorological dataset is used for the MLS data? Is it MERRA? Please state and give a reference.
4.6: as stated above, it think it is necessary to show a plot of this correlation, and maybe explain a bit more, e.g. on what timescales the same correlation can be used. E.g. could you also use it 5 years later (probably not because of N2O trends).
Figure 5: would be good to also plot the passive O3 derived from the tracer relation to N2O, at least for the in-vortex times.
Line 279, 280: I don't see in this plot that ozone variability is linked to chemical processing. E.g. could be transport too. I suggest to drop this sentence.
Figure 6: It would be good to also plot ozone depletion lines from Manney et al. 2020, and Wohltmann et al. 2021 for a better comparison.
Line 289: I question that persistent cold vortex conditions are the main driver here. To me the main driver is activated chlorine and sunlight. The cold vortex conditions are necessary to activate the chlorine, but after that chlorine and sunlight do the job.
Overall, a neat paper using microwave radiometer data for ozone. I support publication after the comments above have been addressed.