the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 Jul 2026
Estimates of Tropospheric Ozone from Nimbus 7 LIMS and TOMS Datasets
Abstract. This work is a diagnostic study of distributions of tropical tropospheric column ozone (TRCO) based on subtracting Limb Infrared Monitor of the Stratosphere (LIMS) Version 6 (V6) Level 3 stratospheric column ozone (SCO) values from Total Ozone Mapping Spectrometer (TOMS) total ozone for several selected days of 1978–1979. The focus of this study is the low latitudes, where the tropopause is near 100 hPa and zonal wave activity is weak. TRCO exhibits a zonal wave-1 feature in the southern tropics with maximum values near 0° longitude during November 1978. Its zonal variation is 70 to 30 Dobson units (DU) and agrees with that of other similar analysis studies. V6 estimates of SCO and calculated TRCO are less reliable in March and May 1979 because those data are affected by temperature biases of order -1.0 to -2.0 K at 30 and 50 hPa; retrieved ozone at those levels may be too large by ~20 % as a result. The LIMS Level 3 sequential estimation algorithm provides smoothed zonal wave ozone amplitudes and reduced zonal variations of SCO. The corresponding calculated TRCO amplitudes are enhanced and not quantitative in those instances. The findings herein should be helpful to others, as they analyze infrared limb measurements of ozone and subtract their stratospheric column amounts from coincident total ozone to obtain accurate zonal variations of tropospheric ozone.
- Preprint
(16508 KB) - Metadata XML
-
Supplement
(11206 KB) - BibTeX
- EndNote
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2026-3796', Anonymous Referee #1, 29 Sep 2026
-
AC1: 'Reply on RC1', Ellis Remsberg, 15 Oct 2026
Author Responses to Comments of Reviewer 1
Opening paragraph:
I appreciate that you carefully reviewed the manuscript. At the end of your opening paragraph, you ask whether there are any direct comparisons of TRCO with observations or climatologies or with ozonesonde data. Remsberg et al. (2013) is one such study, where the LIMS V6 profiles were introduced into the Goddard GEOS-5.7.2 model along with SBUV V8.6 SBUV ozone profiles and the assimilated results were compared with those from a control run using only the SBUV ozone profiles. Section 3.4 of that paper shows that modeled total ozone was improved as compared to TOMS data with the assimilation of the V6 profiles both in the NH and SH and in January and April 1979. However, that improvement is due to differences in the stratospheric column amounts of LIMS versus SBUV and does not characterize the respective estimates of TRCO. While the results from those model runs are no longer available in the Goddard mass storage archive, it is also noted that no tropospheric ozone profiles were included in the two assimilation runs.
Other TRCO climatologies most appropriate to the time of the LIMS data are Fishman et al. (1990) and Ziemke et al. (1998 and references therein). Still, there are very few ozonesonde soundings in the WOUDC data archive from the time of LIMS, especially at low latitudes and in the southern hemisphere. Coincidence is also an issue when making direct comparisons between LIMS and ozonesonde profiles. That is why I preferred to use the LIMS V6 Level 3 (map) product for the present study. I will make the above points more clearly in my revised manuscript.
Major comments:
Although my overall goal for this study is the determination of the latitudinal and zonal variability of TRCO values during the LIMS mission, Section 3 first presents monthly zonal mean values in Table 1 for the latitude range is ±30°, where the tropopause is close to 100 hPa, as in the case of Fishman et al. (1990). Table 1 shows anomalously low values in the SH tropics and subtropics from February through May that cannot be due to random errors in the retrieved ozone profiles. There must be systematic errors leading to those low values, and that is the focus of my investigation in Section 4 and the Supplemental material. While Table 1 is limited to latitudes between ±30, the figures in Section 4 show that the anomalously low values extend to higher latitudes, even though those TRCO calculations also extended from ground to 100 hPa (and not to the tropopause itself which would have given even lower TRCO values). For this reason, I prefer to show the results of Figs. 5, 9, 12 at the higher latitudes, as well. On the other hand, I will re-order the material in Section 4, showing November results first and then March and May, in accordance with Figure 1. Section 4 also indicates that the present use of the LIMS Level 3 coefficients yields zonal variations of TRCO that are not accurate, so I discuss why that may be the case.
Minor comments:
L18-21—This sentence is unrelated to the one before. Instead, I will re-write L16-21 as follows. While the V6 estimates of tropical and subtropical SCO and their calculated TRCO appear reasonable for November 1978, they are less reliable in March and May 1979 because those months have average temperature biases of order -1.0 to -2.0 K at 30 and 50 hPa. V6 ozone in those instances can be too large by ~20%. The LIMS Level 3 sequential estimation algorithm also smooths the zonal wave ozone amplitudes in all months. Zonal variations of V6 SCO are underestimated as a result, although less so in November when zonal wave forcings are weaker.
L99—I will rewrite this to say “the variation of the differences with altitude and latitude in the upper stratosphere for all three months. And yes, Figure 1 is copied from Remsberg et al. (2007, their Figure 8).
L110-112—Table 1 in Remsberg et al. (2007) contains estimates of both random and systematic errors for single retrieved V6 ozone profiles, and the systematic errors dominate. As you note, random errors would not affect monthly average ozone profiles.
L139—I am referring to the lower left panel of Fig. 2, which shows slight negative differences at 30 and 50 hPa; the top left panel shows that the tropopause or altitude of minimum temperatures is at about 100 hPa. In the Supplement Fig. S10 shows V6 zonal mean temperature for 26 January with its approximate tropopause of ~100 hPa, and Fig. S3 indicates that mean V6 temperatures at 20S to 40S are also too cold compared to MERRA at 30 and 50 hPa on that day. I will revise L139-140, accordingly.
L161—Although the altitude of the tropopause was not diagnosed as part of the V6 dataset, a generic definition from the zonal mean temperature distribution would be the pressure-altitude of the temperature minimum (~100 hPa in the tropics and subtropics), which is very similar to the tropopause values in Fishman et al. (1990).
L172-Table 1—I do not know the actual systematic errors associated with the V6 SCO (and consequently its TRCO) values. Instead, I will provide a typical range of zonal TRCO values from Fishman et al. and Ziemke et al. for comparison.
Paragraph L193-L205—I will remove the sentence at L199 that refers to the seasonal changes in total ozone at Natal.
L252—Although TOMS ozone is underestimated slightly in regions of elevated orography, the LIMS measurements also smooth the effects of mountain lee waves on the temperature structure.
L255—Well, I did say, “Notably…at 70°W”, but I will revise the sentence to indicate that there are unphysical values across most longitudes of the SH middle latitudes. I think that is an important reason for showing results at latitudes poleward of ±30°, too.
L255-258—I will raise the issue of effect of temperature bias to earlier in the paragraph, as you suggest. I will also point out that temperature biases vary zonally most likely and that the effects of those biases are non-linear on retrieved ozone.
L277-279—You raise an important point. Although the SE algorithm smooths the zonal wave amplitudes of the V6 ozone and affects the local (in longitude) estimates of TRCO, you are correct in thinking that the SE smoothing should not alter zonal average ozone. Only systematic errors would account for unrealistically low February to May values in Table 1. Note that the upper right panel of Figure S3 (of the Supplement) shows that the standard deviation of temperature about the zonal mean, is quite large (~5K at 70 hPa). If the V6 temperatures have biases, it is more likely that zonal variations of those biases are a primary cause of the zonal variations of the larger negative TRCO values at some SH longitudes. For example, the approximate zonal average TRCO for 15 March at say, 20°S, is clearly negative in Fig. 5 compared with the March mean of 0 DU in Table 1, an indication that the zonal average for 15 March is not representative of the monthly zonal mean value. On the other hand, zonal average TRCO at 20°S on 6 May in Fig. 9 does compare closely with the monthly mean value of -3 DU in Table 1. I will add these comments to the revised manuscript.
Paragraph L281-L288—Figure S3 (lower left panel) shows that there is a clear negative bias in the V6 temperatures below the 100 hPa level, a finding that will affect retrieved ozone at those levels. Even so, the amount of column ozone is small from those lower altitudes, such that a larger bias error in temperature will have only a small effect on SCO (and TRCO). Having said this, I have decided to remove the last paragraph of Section 4a from my revised manuscript. It does not add anything to the main points that I want the reader to take away from my study.
6 May 1979 paragraph—I agree with your range of values. I was focused on the southern tropics. I will make that change in the revised manuscript.
L305—Good catch! I showed regions of missing TOMS measurements as deep blue in an earlier version of Fig. 10, but they are white in my final version. I will correct my misstatement in the revision.
Conclusions section—I will note that I applied the residual method to this early LIMS dataset.
Technical comments:
L43—I will make this change.
L120—I should have said ‘vertical layering in the ozone profiles…’. I am referring to the fact that oftentimes there is vertical structure in ozone, especially near the jet stream, where the LIMS cloud screening algorithm may have flagged those instances as cloud signatures.
L131—Actually both the zonal mean profiles (top left) and their differences (bottom left).
L137—I will make the change.
L178-179—I agree that your transition sentence is better.
L215—I don’t know how I missed that; I’ll make the change.
L218—OK, I’ll delete the phrase.
L219—I’ll make the change and be consistent with its mention several lines later.
L220—I’ll add the modifier.
L236—I’ll check to see if that can be done.
L361—I’ll make the change.
Citation: https://doi.org/10.5194/egusphere-2026-3796-AC1
-
AC1: 'Reply on RC1', Ellis Remsberg, 15 Oct 2026
-
RC2: 'Comment on egusphere-2026-3796', Anonymous Referee #2, 19 Oct 2026
The paper is dedicated to retrievals of tropospheric ozone column from combination of historic limb measurements by LIMS and nadir measurements by TOMS. This is a diagnostic study on application of the well-known residual method to TOMS and LIMS measurements. The target application – tropospheric ozone column – has difficulties due to combination of several factors. However, the paper contains a useful cautionary analysis for users of these satellite records.
This manuscript reads more like a technical report than a scientific paper; it contains illustrative examples but misses systematic analyses and discussion on general methodological/feasibility aspects. Many results are presented for selected days only; it is unclear if they are representative for all days/ for the whole dataset.
The uncertainty analysis for the main target parameters - stratospheric and tropospheric column is not fully characterized (but some components are quantified).
Below are my comments and suggestions on presentation the material and the results.
MAIN COMMENTS
1) In my opinion, it would be good to add more details on motivation for this study In Sect.1. What could be an added value compared to the existing datasets, for example TOMS-CCD, in terms of spatio-temporal resolution and coverage? In this connection, it is worth to present already in the introduction the main parameters of both TOMS and LIMS datasets: the number of data per day, vertical resolution of LIMS, estimated uncertainties.
2) General comment, it will be repeated below in specific comments. Many results are presented for selected days only; it is unclear if they are representative for all days/typical conditions.
3) For tropospheric ozone column, scientists usually use clear-sky nadir measurements, in order to avoid contamination with the ghost column. For your application, do you use clear-sky TOMS total ozone?
4) L.95-105 and Figure 1. More details are needed:
- Is LIMB minus SBUV difference computed using collocated data?
- What are uncertainties for combined differences?
- Are selected days representative for the whole dataset?
- Figure 1 shows increasing LIMS-SBUV differences over time. Using more data could provide a more reliable estimate for these differences. It can be, e.g., temporal evolution of the differences in each month (for one month, there should be sufficient number of collocated observations for reliable bias estimates)
5) Figure 2: Again, MERRA and LIMS temperature comparison is presented for one day only. Is the structure of biases the same for all days?
6) Section 3. Due to different sampling patterns, the difference between monthly mean TOMS total column and monthly mean LIMB stratospheric column has also a sampling-related component. I suggest indicating in the text that this analysis has been performed for illustration of compatibility of TOMS total column and LIMS stratospheric column. Maybe, even worth to rename this section.
This analysis would be more solid, if the difference were evaluated on collocated data (and then averaged in monthly values) and with estimated uncertainties.
7) Lines 230-245 on horizontal interpolation and SCO.
- Since lower stratosphere has dynamical perturbation, computing first stratospheric column and then applying interpolation would be more accurate
- Using Fourier coefficients in zonal wavenumber can introduce significant interpolation errors in the regions with pronounced gradients and localized perturbations. The discussion in Lines 268-277 on the limitations of the SE algorithm should be presented together with the description of horizontal interpolation in lines 230-245.
- There are more advanced interpolation methods, for example, kriging.
8) I suggest presenting Figures 3-5 in one figure with 3 panels: this will help to see all three columns in the same plot. Since the lower limit of ~100 hPa for stratosphere/troposphere separation, SCO outside the tropics contains only a part of stratospheric column. Therefore, I suggest either removing SCO and TRCO outside ~30S - 30N or shading this region.
The same suggestion for combination of Figures 7-9 and Figure 10-12.
9) Illustrations and discussion for 3 days in November, March and May are interesting, but their representativeness remains unclear. Are all days in March similar to 15 March? This would help to separate the systematic component from day-to-day variations.
What are estimated uncertainties of TRCO for individual days?
Note that even for modern datasets like OMI-MLS or OMPS-MERRA2 the tropospheric ozone is provided as monthly mean data.
9) November 1978 TRCO demonstration. I suggest to add monthly mean map (lat-lon) of tropospheric ozone: then the morphology of tropospheric ozone could be compared with the modern climatology of tropospheric ozone, and the hypothesis on the features related to biomass burning will be better justified.
10) Since the TRCO estimates for November 1978 look reasonable (interesting that it is for the very beginning of both LIMS and TOMS missions), a general discussion on improving compatibility of TOMS and LIMS would be useful in “Conclusion” section. For example, can using MERRA temperatures in LIMS retrievals decrease LIMS biases at lower altitudes? What can be the reason for progressing biases and how this can be corrected?, etc.
MINOR COMMENTS
L.16: “V6” -> “LIMS V6”
L.97: Is the vertical resolution of 3.7 km for all altitudes/latitudes/seasons?
L.124 “Climatological ozone values” - From which climatology?
L.129-131: “Note that their zonal standard deviations about the respective zonal means agree (top right), an indication that both data sets are equally responsive to zonal temperature waves.”
Yes, but they strongly disagree in the UTLS
Citation: https://doi.org/10.5194/egusphere-2026-3796-RC2
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2026-3796', Anonymous Referee #1, 29 Sep 2026
This manuscript presents an exploratory attempt to combine LIMS and TOMS stratospheric and total ozone columns, respectively, to obtain tropospheric ozone column (TRCO) data. The LIMS observations are among the first limb measurements ever performed from a satellite platform, and this study investigates the application of a residual technique to these very early observations. This aspect makes the study interesting; however, the scientific value of the resulting data is quite limited, both because of the short time span covered by the data set and because of the sub-optimal quality of the available observations: as the author states at the end, reasonable values of TRCO are possible to obtain only for November 1978. The author provides quite extensive explanations of possible causes for the obtained unrealistic TRCO values in the other analyzed periods. However, TRCO values are not directly inter-compared with other observations or a climatology: is there any profile data that can be integrated and used for comparison in the troposphere? No ozonesonde available at that time? Comments are detailed in the following.
Major comments
- A wish that I think would improve the paper, is about the comparison with tropospheric other ozone values, if available. Is there any ozonesonde available for November 1978 that can be used for comparison? If not available, I suggest to state it in the paper. Alternatively, the good results obtained in Nov 1978 could be compared with a November climatology to have an impression on how the retrieved results compare.
- In the abstract the author says that the focus of the study is in the tropics, but then he provides the TRCO results on global maps (Fig.5, 9, 12) even though for March and May 1979 especially the outer tropical values are clearly nonphysical. I suggest to limit the latitude span of these plots, especially Fig. 9 and 12 to ±30°. I am also wondering if it would be better to show first Nov 1978 that works nicely and afterwards the cases that are not working and discuss the possible causes.
- A methodological question: the paper presents a L3 residual approach, would it be possible or has it been tested a residual calculation for collocated TOMS/LIMS observations? Even a few cases could give a hint if this way gives useful information or the discussed biases prevent getting reasonable results.
- Overall, I find that the different parts of the paper are not well connected. Questions are posed to guide the reader in the investigations but I suggest a revision of the paper including some more connections between the paragraphs within sections and between sections. For example,
Minor comments
- L18-21: I find this sentence a bit confusing here, is it related to the previous sentence on the temperature bias? As a reader of the abstract, it is not clear if this is happening only for March and May 1979 or is overall the case?
- L99: It is not clear to me which pattern the author is referring to, as a function of latitude or across the three pictures? In addition, the quality of Fig.1 is poor; can it be improved? Is this taken from an older paper?
- L110-112: Are these errors random or systematic? You mention ±1.2 K, random, shouldn’t this be reduced in magnitude with averaging on a monthly basis?
- L139: In the Supplements there is only a plot with the SH profiles, can you add the figure number you are referring to? It is also not clear at which altitude the author is referring to. Do you have the tropical tropopause height for that day that you could include?
- L161: Can you be more specific about the tropopause? Which definition are you using? This information could also be used in Fig.1 to visualize the lower bottom of SCO on average.
- L172-Table 1: Would it be possible to add an estimated uncertainty to these values? For example in terms of combined std of the monthly SCO and TCO used to compute the residual, to have an idea of the variability of the used quantities. Alternatively, could you add a typical range of zonal TRCO, for example from Ziemke et al., at least for the tropics?
- Paragraph L193-L205: You say that TRCO data from Ziemke’s papers show a decrease of ~20 DU from Nov to May, this justifies the decrease of 30 DU (or 20 DU after the adjustment) you see at 5° S. I would remove the Dobson observation from this reasoning as it might be related to the specific day considered, whereas satellite data are monthly averages, so not directly comparable.
- L252: “or where there are decreases…” not clear, do you mean that you selected location where the TRCO is lower due to orography? Why is this choice made?
- L255: Nonphysical values are found not only at 70° W but overall in the SH.
- L255-258: Can you please state from the beginning that the temperature bias can explain only partially the negative values, not only at the very end?
- L277-279: The author writes that “that zonal variations of TRCO are qualitative at best”. Do you then expect that once zonally averaged the TRCO results should be more realistic, taking into account the instrumental biases you mention? It seems to me that zonal averages in SH tropics are also generally negative for this day, regardless of their longitude variations.
- Paragraph L281-L288: The author seems to suggest the extension of the dataset at lower levels but then concludes that it is not reliable, so that “TRCO distributions are only evaluated for the tropics and subtropics”. However the next pictures are still showing mid-latitude regions.
- 6 May 1979 paragraph. In the tropics (20°S to 20°N) I see values between 80 and -40 DU, not from 0 to
- L305: I don’t see deep blue areas in Fig.10 at northern high latitudes.
- I would state also in the conclusion that this is the first case of application of a residual technique to satellite observations, which make this study interesting.
Technical comments
- L43: Please add “In the following,” at the beginning of the new sentence.
- L120: What do you mean with “a vertical layering of the ozone profiles near the tropopause”?
- L131: “bottom left” → “top left”.
- L137: “or” → “i.e.,”
- L178-179: I suggest to rewrite the last sentence like “Possible causes of these unrealistic features in March and May are discussed in the next paragraphs.”
- L215: “20° W” → “50° W”.
- L218: I would remove “from west to east longitude”.
- L219: “Poona” → “Pune”.
- L220: Please add “Collocated” before “Dobson/TOMS tropical total ozone”.
- L236: I would put this details in a footnote.
- L361: “not as reliable” → “not reliable”.
Citation: https://doi.org/10.5194/egusphere-2026-3796-RC1 -
AC1: 'Reply on RC1', Ellis Remsberg, 15 Oct 2026
Author Responses to Comments of Reviewer 1
Opening paragraph:
I appreciate that you carefully reviewed the manuscript. At the end of your opening paragraph, you ask whether there are any direct comparisons of TRCO with observations or climatologies or with ozonesonde data. Remsberg et al. (2013) is one such study, where the LIMS V6 profiles were introduced into the Goddard GEOS-5.7.2 model along with SBUV V8.6 SBUV ozone profiles and the assimilated results were compared with those from a control run using only the SBUV ozone profiles. Section 3.4 of that paper shows that modeled total ozone was improved as compared to TOMS data with the assimilation of the V6 profiles both in the NH and SH and in January and April 1979. However, that improvement is due to differences in the stratospheric column amounts of LIMS versus SBUV and does not characterize the respective estimates of TRCO. While the results from those model runs are no longer available in the Goddard mass storage archive, it is also noted that no tropospheric ozone profiles were included in the two assimilation runs.
Other TRCO climatologies most appropriate to the time of the LIMS data are Fishman et al. (1990) and Ziemke et al. (1998 and references therein). Still, there are very few ozonesonde soundings in the WOUDC data archive from the time of LIMS, especially at low latitudes and in the southern hemisphere. Coincidence is also an issue when making direct comparisons between LIMS and ozonesonde profiles. That is why I preferred to use the LIMS V6 Level 3 (map) product for the present study. I will make the above points more clearly in my revised manuscript.
Major comments:
Although my overall goal for this study is the determination of the latitudinal and zonal variability of TRCO values during the LIMS mission, Section 3 first presents monthly zonal mean values in Table 1 for the latitude range is ±30°, where the tropopause is close to 100 hPa, as in the case of Fishman et al. (1990). Table 1 shows anomalously low values in the SH tropics and subtropics from February through May that cannot be due to random errors in the retrieved ozone profiles. There must be systematic errors leading to those low values, and that is the focus of my investigation in Section 4 and the Supplemental material. While Table 1 is limited to latitudes between ±30, the figures in Section 4 show that the anomalously low values extend to higher latitudes, even though those TRCO calculations also extended from ground to 100 hPa (and not to the tropopause itself which would have given even lower TRCO values). For this reason, I prefer to show the results of Figs. 5, 9, 12 at the higher latitudes, as well. On the other hand, I will re-order the material in Section 4, showing November results first and then March and May, in accordance with Figure 1. Section 4 also indicates that the present use of the LIMS Level 3 coefficients yields zonal variations of TRCO that are not accurate, so I discuss why that may be the case.
Minor comments:
L18-21—This sentence is unrelated to the one before. Instead, I will re-write L16-21 as follows. While the V6 estimates of tropical and subtropical SCO and their calculated TRCO appear reasonable for November 1978, they are less reliable in March and May 1979 because those months have average temperature biases of order -1.0 to -2.0 K at 30 and 50 hPa. V6 ozone in those instances can be too large by ~20%. The LIMS Level 3 sequential estimation algorithm also smooths the zonal wave ozone amplitudes in all months. Zonal variations of V6 SCO are underestimated as a result, although less so in November when zonal wave forcings are weaker.
L99—I will rewrite this to say “the variation of the differences with altitude and latitude in the upper stratosphere for all three months. And yes, Figure 1 is copied from Remsberg et al. (2007, their Figure 8).
L110-112—Table 1 in Remsberg et al. (2007) contains estimates of both random and systematic errors for single retrieved V6 ozone profiles, and the systematic errors dominate. As you note, random errors would not affect monthly average ozone profiles.
L139—I am referring to the lower left panel of Fig. 2, which shows slight negative differences at 30 and 50 hPa; the top left panel shows that the tropopause or altitude of minimum temperatures is at about 100 hPa. In the Supplement Fig. S10 shows V6 zonal mean temperature for 26 January with its approximate tropopause of ~100 hPa, and Fig. S3 indicates that mean V6 temperatures at 20S to 40S are also too cold compared to MERRA at 30 and 50 hPa on that day. I will revise L139-140, accordingly.
L161—Although the altitude of the tropopause was not diagnosed as part of the V6 dataset, a generic definition from the zonal mean temperature distribution would be the pressure-altitude of the temperature minimum (~100 hPa in the tropics and subtropics), which is very similar to the tropopause values in Fishman et al. (1990).
L172-Table 1—I do not know the actual systematic errors associated with the V6 SCO (and consequently its TRCO) values. Instead, I will provide a typical range of zonal TRCO values from Fishman et al. and Ziemke et al. for comparison.
Paragraph L193-L205—I will remove the sentence at L199 that refers to the seasonal changes in total ozone at Natal.
L252—Although TOMS ozone is underestimated slightly in regions of elevated orography, the LIMS measurements also smooth the effects of mountain lee waves on the temperature structure.
L255—Well, I did say, “Notably…at 70°W”, but I will revise the sentence to indicate that there are unphysical values across most longitudes of the SH middle latitudes. I think that is an important reason for showing results at latitudes poleward of ±30°, too.
L255-258—I will raise the issue of effect of temperature bias to earlier in the paragraph, as you suggest. I will also point out that temperature biases vary zonally most likely and that the effects of those biases are non-linear on retrieved ozone.
L277-279—You raise an important point. Although the SE algorithm smooths the zonal wave amplitudes of the V6 ozone and affects the local (in longitude) estimates of TRCO, you are correct in thinking that the SE smoothing should not alter zonal average ozone. Only systematic errors would account for unrealistically low February to May values in Table 1. Note that the upper right panel of Figure S3 (of the Supplement) shows that the standard deviation of temperature about the zonal mean, is quite large (~5K at 70 hPa). If the V6 temperatures have biases, it is more likely that zonal variations of those biases are a primary cause of the zonal variations of the larger negative TRCO values at some SH longitudes. For example, the approximate zonal average TRCO for 15 March at say, 20°S, is clearly negative in Fig. 5 compared with the March mean of 0 DU in Table 1, an indication that the zonal average for 15 March is not representative of the monthly zonal mean value. On the other hand, zonal average TRCO at 20°S on 6 May in Fig. 9 does compare closely with the monthly mean value of -3 DU in Table 1. I will add these comments to the revised manuscript.
Paragraph L281-L288—Figure S3 (lower left panel) shows that there is a clear negative bias in the V6 temperatures below the 100 hPa level, a finding that will affect retrieved ozone at those levels. Even so, the amount of column ozone is small from those lower altitudes, such that a larger bias error in temperature will have only a small effect on SCO (and TRCO). Having said this, I have decided to remove the last paragraph of Section 4a from my revised manuscript. It does not add anything to the main points that I want the reader to take away from my study.
6 May 1979 paragraph—I agree with your range of values. I was focused on the southern tropics. I will make that change in the revised manuscript.
L305—Good catch! I showed regions of missing TOMS measurements as deep blue in an earlier version of Fig. 10, but they are white in my final version. I will correct my misstatement in the revision.
Conclusions section—I will note that I applied the residual method to this early LIMS dataset.
Technical comments:
L43—I will make this change.
L120—I should have said ‘vertical layering in the ozone profiles…’. I am referring to the fact that oftentimes there is vertical structure in ozone, especially near the jet stream, where the LIMS cloud screening algorithm may have flagged those instances as cloud signatures.
L131—Actually both the zonal mean profiles (top left) and their differences (bottom left).
L137—I will make the change.
L178-179—I agree that your transition sentence is better.
L215—I don’t know how I missed that; I’ll make the change.
L218—OK, I’ll delete the phrase.
L219—I’ll make the change and be consistent with its mention several lines later.
L220—I’ll add the modifier.
L236—I’ll check to see if that can be done.
L361—I’ll make the change.
Citation: https://doi.org/10.5194/egusphere-2026-3796-AC1
-
RC2: 'Comment on egusphere-2026-3796', Anonymous Referee #2, 19 Oct 2026
The paper is dedicated to retrievals of tropospheric ozone column from combination of historic limb measurements by LIMS and nadir measurements by TOMS. This is a diagnostic study on application of the well-known residual method to TOMS and LIMS measurements. The target application – tropospheric ozone column – has difficulties due to combination of several factors. However, the paper contains a useful cautionary analysis for users of these satellite records.
This manuscript reads more like a technical report than a scientific paper; it contains illustrative examples but misses systematic analyses and discussion on general methodological/feasibility aspects. Many results are presented for selected days only; it is unclear if they are representative for all days/ for the whole dataset.
The uncertainty analysis for the main target parameters - stratospheric and tropospheric column is not fully characterized (but some components are quantified).
Below are my comments and suggestions on presentation the material and the results.
MAIN COMMENTS
1) In my opinion, it would be good to add more details on motivation for this study In Sect.1. What could be an added value compared to the existing datasets, for example TOMS-CCD, in terms of spatio-temporal resolution and coverage? In this connection, it is worth to present already in the introduction the main parameters of both TOMS and LIMS datasets: the number of data per day, vertical resolution of LIMS, estimated uncertainties.
2) General comment, it will be repeated below in specific comments. Many results are presented for selected days only; it is unclear if they are representative for all days/typical conditions.
3) For tropospheric ozone column, scientists usually use clear-sky nadir measurements, in order to avoid contamination with the ghost column. For your application, do you use clear-sky TOMS total ozone?
4) L.95-105 and Figure 1. More details are needed:
- Is LIMB minus SBUV difference computed using collocated data?
- What are uncertainties for combined differences?
- Are selected days representative for the whole dataset?
- Figure 1 shows increasing LIMS-SBUV differences over time. Using more data could provide a more reliable estimate for these differences. It can be, e.g., temporal evolution of the differences in each month (for one month, there should be sufficient number of collocated observations for reliable bias estimates)
5) Figure 2: Again, MERRA and LIMS temperature comparison is presented for one day only. Is the structure of biases the same for all days?
6) Section 3. Due to different sampling patterns, the difference between monthly mean TOMS total column and monthly mean LIMB stratospheric column has also a sampling-related component. I suggest indicating in the text that this analysis has been performed for illustration of compatibility of TOMS total column and LIMS stratospheric column. Maybe, even worth to rename this section.
This analysis would be more solid, if the difference were evaluated on collocated data (and then averaged in monthly values) and with estimated uncertainties.
7) Lines 230-245 on horizontal interpolation and SCO.
- Since lower stratosphere has dynamical perturbation, computing first stratospheric column and then applying interpolation would be more accurate
- Using Fourier coefficients in zonal wavenumber can introduce significant interpolation errors in the regions with pronounced gradients and localized perturbations. The discussion in Lines 268-277 on the limitations of the SE algorithm should be presented together with the description of horizontal interpolation in lines 230-245.
- There are more advanced interpolation methods, for example, kriging.
8) I suggest presenting Figures 3-5 in one figure with 3 panels: this will help to see all three columns in the same plot. Since the lower limit of ~100 hPa for stratosphere/troposphere separation, SCO outside the tropics contains only a part of stratospheric column. Therefore, I suggest either removing SCO and TRCO outside ~30S - 30N or shading this region.
The same suggestion for combination of Figures 7-9 and Figure 10-12.
9) Illustrations and discussion for 3 days in November, March and May are interesting, but their representativeness remains unclear. Are all days in March similar to 15 March? This would help to separate the systematic component from day-to-day variations.
What are estimated uncertainties of TRCO for individual days?
Note that even for modern datasets like OMI-MLS or OMPS-MERRA2 the tropospheric ozone is provided as monthly mean data.
9) November 1978 TRCO demonstration. I suggest to add monthly mean map (lat-lon) of tropospheric ozone: then the morphology of tropospheric ozone could be compared with the modern climatology of tropospheric ozone, and the hypothesis on the features related to biomass burning will be better justified.
10) Since the TRCO estimates for November 1978 look reasonable (interesting that it is for the very beginning of both LIMS and TOMS missions), a general discussion on improving compatibility of TOMS and LIMS would be useful in “Conclusion” section. For example, can using MERRA temperatures in LIMS retrievals decrease LIMS biases at lower altitudes? What can be the reason for progressing biases and how this can be corrected?, etc.
MINOR COMMENTS
L.16: “V6” -> “LIMS V6”
L.97: Is the vertical resolution of 3.7 km for all altitudes/latitudes/seasons?
L.124 “Climatological ozone values” - From which climatology?
L.129-131: “Note that their zonal standard deviations about the respective zonal means agree (top right), an indication that both data sets are equally responsive to zonal temperature waves.”
Yes, but they strongly disagree in the UTLS
Citation: https://doi.org/10.5194/egusphere-2026-3796-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 166 | 102 | 41 | 309 | 98 | 29 | 27 |
- HTML: 166
- PDF: 102
- XML: 41
- Total: 309
- Supplement: 98
- BibTeX: 29
- EndNote: 27
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This manuscript presents an exploratory attempt to combine LIMS and TOMS stratospheric and total ozone columns, respectively, to obtain tropospheric ozone column (TRCO) data. The LIMS observations are among the first limb measurements ever performed from a satellite platform, and this study investigates the application of a residual technique to these very early observations. This aspect makes the study interesting; however, the scientific value of the resulting data is quite limited, both because of the short time span covered by the data set and because of the sub-optimal quality of the available observations: as the author states at the end, reasonable values of TRCO are possible to obtain only for November 1978. The author provides quite extensive explanations of possible causes for the obtained unrealistic TRCO values in the other analyzed periods. However, TRCO values are not directly inter-compared with other observations or a climatology: is there any profile data that can be integrated and used for comparison in the troposphere? No ozonesonde available at that time? Comments are detailed in the following.
Major comments
Minor comments
Technical comments