the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 10 Sep 2026
The ATMOSFER campaign: synergistic observations for the preparation of the ESA FORUM mission
Abstract. The balloon-borne ATMOSFER campaign, supported by CNES, was held in Kiruna in June 2024. This campaign involved lightweight hygrometers (Pico-Light H2O, NOAA GML frost point hygrometer, EN-SCI Cryogenic frost point hygrometer), optical particle counters and ice crystal imager (POPS, LOAC v1.5, LPC and NIXE-B), ozone sondes and the airborne demonstrator of the future ESA FORUM mission developed by CNR-INO (FIRMOS-B instrument). The instruments have been flown aboard Zero Pressure Balloon (ZPB) and small weather balloons (RS) by CNES and the scientific teams. The synergistic observations aboard stratospheric balloons captured the signature of multiple influences which are useful for preparation activities of the ESA FORUM mission. The measurements revealed stratospheric intrusions of sub-tropical air, the presence of upper tropospheric ice cirrus clouds and aerosol layers of small particles potentially originating from volcanic eruption (between 15 and 17 km of altitude) and forest fires (near the polar cold point tropopause), which will influence the vertical profile retrievals from FIRMOS-B radiances and provide interesting features for radiative closure studies. This article aims to give an overview of the observation potential of the campaign for further studies in the frame of the ESA FORUM mission. The datasets are publicly available in the HEMERA datacenter (https://data.hemera-h2020.eu/atmospheric-balloon-experiments/#/).
- Preprint
(2531 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on essd-2026-557', Anonymous Referee #1, 07 Oct 2026
-
RC2: 'Comment on essd-2026-557', Anonymous Referee #2, 08 Oct 2026
Review of ‘The ATMOSFER campaign: synergistic observations for the preparation of the ESA FORUM mission’ by M. Ghysels and coauthors, submitted to ESSD
This manuscript discusses the ATMOSFER balloon campaign based in Kiruna, Sweden that took place in June 2024. The campaign included the launch of a large balloon (and coincident, smaller balloon deployments) capable of deploying a large range of instruments that observe ice microphysics, aerosols, ozone, water vapor, and high-spectral-resolution mid to far infrared radiances by the FIRMOS-B instrument. The FIRMOS-B instrument is a pathfinder in preparation for ESA’s FORUM mission, that will spectrally resolve the MIR and FIR simultaneously (in synergy with IASI spectra). The mission overview discusses the instruments, observing strategies, and presents initial results of the instrument measurements, with some initial interpretation with respect to model back trajectories near fire events in North America. Shown are initial results of temperature and moisture retrievals from FIRMOS-B that appear to be promising. However, most of the FIRMOS-B retrieval analysis of cloud properties will be reported elsewhere in a separate, future manuscript (justifiably so).
I found this paper to be an enjoyable read and mostly organized well. There are some places that could use further clarification and improvements in the figures could be made. Overall, my suggestions are quite minor, anticipating that this field campaign data will be a valuable resource for other satellite missions beyond FORUM algorithm development and testing.
Lines 70-72: Technically FORUM will not be the first to cover the FIR component. The PREFIRE CubeSats have made spectrally resolved global observations for over two years. However, the spectra extend only up to ~50 microns, mixing windows and absorption lines together, and do not offer the high spectral resolution capabilities afforded by FORUM. Some rewrite to make lines 70-72 more precise is warranted.
Lines 96-98: Regarding the one spectrum per minute sampling, how does this work out to a ground footprint size and spatial sampling with respect to time? As the balloon platform might drift a bit, does the ground footprint smear more with faster movement because the instrument integrates the signal over time? How about changes in the footprint size with altitude as it will get smaller from the surface to the upper troposphere? Is the pointing of FIRMOS-B held steady or controlled, or does it wobble? While a more detailed paper will follow, some additional details in the instrument description are warranted.
Section 3: I was quite confused between the different balloon launches, and which instruments were staged on each of the different platforms. Figure 4 shows some graphics of the different configurations. Perhaps additional data tables of what instrument was on which balloon, when each balloon was launched, and other metadata would be helpful to make clearer what was launched when.
Line 229: Is NIXE-B “lower” or “downstream”? Is downstream common terminology for balloon campaigns (this is asked by a reviewer with no expertise in balloon campaigns)?
Figure 5 could use an addition on the legend for the green line. Also, in the caption, should be “is” not “if”.
Lines 307-311: How close are these other balloons to the main gondola in space and time? Would be interesting to see how close horizontally they all are to each other.
Line 355: ‘percent’
For Figure 8, is there any reason to show the whole polar region? Perhaps there is some connection with the fires in Canada but it wasn’t discussed explicitly in the manuscript. I found these three panels as is a bit of a misfit in terms of how they are used to set the stage of the campaign.
Figure 9 is interesting, but it needs a lot of work. The quality isn’t great. The y-axes for altitude need to be lined up. The green dots for the large NIXE-B sizes are hard to see and are obscured by the red dashed line. The green dots in the middle panel for (I think) the LOAC are impossible to sort from the other dots. This is a key figure and connects the synergy of the other instruments.
Figure 11 is an average cloud top height over an hour-long period? Why present it that way? What about three different times, one each on + and - 30 min from the balloon, and one centered on the balloon? I’m also a bit confused because the balloon observes over many hours while this is a one-hour average. Some clarification is needed.
Figure 12 is interesting but I can’t read the times. Also, why are so many trajectories needed?
Lines 466-467: I didn’t see figures for the ACE-FTS and SAGE III extinction profiles. Those would be helpful to have.
Line 480: Since there is quite a bit of tropospheric and stratospheric nuance to this campaign, the ozone discussion seems lacking. Ozone should be very elevated in the stratosphere, but it appears at this line a tropospheric portion of the ozone profile is discussed?
Line 565: IWC is mentioned (but not shown – to appear elsewhere). Is this a function of height? What about effective radius? How can one get these as a function of height reliably from a high spectral resolution IR sounder? References would be appreciated.
Citation: https://doi.org/10.5194/essd-2026-557-RC2 -
RC3: 'Comment on essd-2026-557', Anonymous Referee #3, 09 Oct 2026
The manuscript presents the ATMOSFER balloon campaign carried out in Kiruna in June 2024. The combination of far-infrared radiance measurements, water vapour observations from different hygrometers, ozone profiles, aerosol measurements and ice-particle imaging makes this campaign particularly interesting. The dataset has clear potential for the preparation of the FORUM mission and also for other studies of the upper troposphere and lower stratosphere.
I appreciate the considerable experimental work behind this campaign, including the deployment of new instrumental configurations. I also appreciate that the authors discuss several difficulties encountered during the measurements.
I consider the paper suitable for ESSD, but some aspects need clarification before publication. In particular, I found inconsistencies in the description of the instruments and flights, problems in the statistical equations, and insufficient information on some processing and quality-control procedures. In a few places, in my opinion the conclusions appear stronger than the evidence presented.
At the present stage the authors do not need to include a complete radiative closure study in this paper. However, the reader should be able to understand which measurements can be used, with which uncertainties, and how the main comparisons were performed. For these reasons, I recommend a revision of the manuscript.
Major comments
1. Instruments and flight chronology
I had some difficulty understanding which version of LOAC was used during the campaign. Section 2.5 describes LOAC-2, while the abstract, the gondola description, several figures and the data-availability table refer to LOAC v1.5. Could the authors clarify this point and check that the instrument characteristics given in the paper correspond to the version actually flown?
There is also an inconsistency concerning NIXE-B. Lines 238–239 indicate that it was flown with CFH during the second session, whereas Section 4.3.1 and Figure 10 assign the free-balloon measurements to the first session. The statement that no aerosol measurements were made during the second session also needs to be reconciled with the instrument configuration described earlier.
I think that a single table summarising all flights would be very useful. It could include the actual launch times, instruments and configurations, measurement periods, valid-data intervals and corresponding dataset identifiers. Launch times should be distinguished from the starting times of the archived observations.
2. Statistical equations
Equations (2) and (3) should be checked carefully. As written, Equation (2) simplifies to the sum of the group means, because nₖ appears in both the numerator and denominator of each term. For a weighted mean, the denominator should instead be the sum of the group sizes.
Equation (3) also appears unclear: it is not clear whether the authors intend to calculate a pooled within-group standard deviation or the standard deviation of all observations considered together. These quantities have different meanings.
This distinction is important here because the differences between sessions can be large and can change sign. For example, Table 2 reports stratospheric mean differences of +12% and −16.5%. If the purpose is to describe the overall dispersion, the difference between the session means must also be included.
I would ask the authors to correct the equations, provide the number of observations and weights used, and check the combined statistics in Tables 1–3. Please also clarify whether these are only errors in the written equations or whether the numerical calculations are affected. The meaning of the values following “±” should be stated explicitly.
3. Description of the water vapour comparisons
Some details of the comparison procedure are not consistent. Lines 313–315 mention 150 m bins, while lines 322–326 refer to intervals of ±100 m. Equation (1) uses FPH as reference, but the convention adopted for the Pico-Light–CFH comparison is not specified.
The altitude intervals also need checking. For the first-session FPH–CFH stratospheric comparison, the text mentions 22–24 km, whereas Table 3 gives much wider intervals. The tropospheric summary at lines 353–355 does not seem to agree with Table 1.
Could the authors give a more explicit description of the procedure, including the vertical grid, matching of observations, averaging, and rejection of invalid measurements? For each comparison, the actual altitude range and number of valid pairs would be helpful. Particular attention should be given to the treatment of CFH measurements affected by contamination or unstable operation.
4. Spatial and temporal separation of the measurements
The hygrometers were flown on different balloons. Although the launches were close in time, the instruments did not necessarily measure the same air masses, particularly during descent. In my view, this deserves more attention in the interpretation of the comparisons. Near the tropopause, where the paper reports strong gradients, even a relatively small displacement of atmospheric layers may produce appreciable differences between profiles.
The explanations proposed for the FPH discrepancies, involving the response of the control system, are reasonable possibilities. However, they would be more convincing if supported by instrumental diagnostics. Improved agreement after excluding part of a profile does not, by itself, demonstrate instrumental reproducibility.
I suggest showing the balloon trajectories and reporting the spatial and temporal separation of the observations as a function of altitude. This would possibly help the reader distinguish instrumental differences from atmospheric variability. Where data are excluded, the criteria should be based as far as possible on independent quality information.
5. Performance of the modified hygrometers
The paper describes contamination, difficulties in stabilising the CFH controller and descent speeds reaching approximately 60 m s⁻¹. In the second session, the stratospheric differences relative to FPH reach about 37%.
Under these circumstances, I do not think that the uncertainty generally reported for CFH can be directly applied to the liquid-nitrogen prototype used here. A similar distinction is needed between the established performance of FPH and the performance of its dry-ice/alcohol configuration during these flights.
It would be useful to state more clearly which parts of each profile are considered reliable for quantitative use, which are uncertain, and which should be excluded, and justify such choices. These distinctions should also be present as quality flags in the distributed files.
6. Aerosol and ice-particle measurements
The availability of several complementary particle instruments is one of the strengths of the campaign. However, I would welcome more information on the comparability of their products.
Please clarify the size definitions, calibration assumptions and refractive indices used, together with sampling volumes, flow corrections, collection efficiencies and counting uncertainties. The distinction between optical-equivalent diameter and area-equivalent diameter from images is especially important.
The normalisation of the distributions in Figure 9 is not completely clear to me, particularly the expression “accumulated concentration per 100 m interval”. A quantitative comparison in overlapping size ranges, where the instruments permit it, would strengthen this part of the paper.
The calculation of IWC at lines 470–476 also needs further explanation. Which mass–dimension relationships were used? How was particle shape considered? Were both NIXE-B modules included, and over which size intervals? Some estimate or discussion of the resulting uncertainty is necessary.
For Figure 10, I suggest reporting the number of particles used to derive the shape percentages and briefly describing the classification procedure.
7. FIRMOS-B radiometric assessment
The comparison with two retrieval codes is useful. Nevertheless, in my opinion, this analysis should be presented more cautiously as a consistency assessment, rather than as a fully independent validation of instrumental accuracy. The spectra are selected using a final χ² below 2, and the residuals of the selected spectra are then used to assess NESR. This selection may favour good agreement. Could the authors define the χ² statistic, indicate how many of the 200 spectra were retained, and explain whether clear-sky conditions were identified independently?
Also, the retrieved atmospheric, surface and instrumental parameters may compensate for part of the measurement or model errors. This possibility should be discussed when interpreting the residuals.
I am not an expert on that, but the assumption that the radiative-transfer models introduce no relevant error seems rather strong. Moreover, previous validation mainly above 650 cm⁻¹ does not necessarily establish the same performance throughout the far-infrared region.
Finally, in Figure 16, some residual features appear to exceed the ARA envelope. I would therefore suggest qualifying the statement at lines 558–560 and specifying the spectral regions or statistical criterion for which the agreement is considered satisfactory.
A clearer distinction between the instrumental uncertainty budget and the information obtained from the retrieval residuals would, in my view, improve this section considerably.
8. Comparison between retrieved profiles and in situ observations
Figure 17 is interesting, but additional information is needed to interpret the comparison. In particular, the vertical resolution of the retrieval, its dependence on the a priori, and the spatial and temporal correspondence with the balloon profiles should be described.
If available, averaging kernels would be useful. Otherwise, an equivalent description of the vertical sensitivity would help. Please also explain the plotted uncertainties and clarify whether the ERA5 profile is used as an a priori constraint, an initial guess, or both.
For a quantitative comparison, the different vertical resolutions should be taken into account. If this is beyond the scope of the present work, I think it is acceptable to retain the figure as a preliminary illustration, provided that this limitation is clearly stated.
The possible influence of aerosols and thin cirrus on the discrepancies near 12 and 16 km is an interesting hypothesis. However, other explanations, including limited retrieval sensitivity, a priori influence and differences in the sampled air masses, should also be mentioned.
9. Interpretation of cloud formation and aerosol origin
I suggest reconsidering the statement at line 472 that the observed RHi values exclude homogeneous freezing. The humidity measured in an already existing cloud is not necessarily the humidity at the time of ice formation. Subsequent crystal growth may have reduced the supersaturation. In my opinion, these observations alone cannot exclude a previous contribution from homogeneous freezing.
Some of the source attributions would also benefit from more cautious wording. A backward trajectory passing over a region with fires or smoke does not necessarily demonstrate contact with the plume at the relevant altitude. Similarly, tropical air-mass origin does not by itself establish a volcanic source for the aerosol, and the absence of an ozone feature does not uniquely identify a high-latitude origin.
Could the authors distinguish more explicitly between what is directly observed and what is proposed as a possible interpretation? Additional details on the trajectory calculations and satellite products would be useful, including initialisation times, vertical-motion assumptions, uncertainties and quality criteria. The possible volcanic origin mentioned in the abstract should either be supported more clearly or expressed with greater caution.
Minor comments
1. Lines 277–286: the ice-saturation mixing ratio is calculated from ambient temperature and pressure, while measured water vapour is used to derive RHi. Please revise the description accordingly and check the reference “Goff Gratch (1984)”, which I could not find in the bibliography.
2. Please avoid describing the whole RHi interval of 80–130% as ice-saturated. Subsaturated and supersaturated conditions should be distinguished.
3. Lines 178–179: could the authors explain how the stated POPS vertical resolution of approximately 100 m is related to the 1 s sampling interval?
4. Lines 530–535: the distinction between the wavenumber offset in the ISRF and the nominal channel wavenumber should be made explicit. Please also define the sinc convention and response normalisation.
5. Figures 14–15: “Epoch” needs units and a time reference. UTC time would probably be easier for the reader.
6. Some labels in Figures 9 and 12 are difficult to read. Figure 12, in particular, would benefit from fewer overlapping trajectory labels.
7. Please identify the Sentinel-5P products and versions used, with the relevant quality criteria and uncertainty information.
8. The bibliography should be checked for missing references, including Krauchi et al. (2016), and completed with the dataset citations. Duplicated DOI prefixes and residual HTML characters should also be corrected.
I consider this a worthwhile contribution, based on a demanding and scientifically useful campaign. My recommendation for major revision mainly reflects the need to make the dataset description and quality assessment more consistent and transparent. I believe that addressing the points above would considerably strengthen the paper and make the observations easier to use with confidence.
Citation: https://doi.org/10.5194/essd-2026-557-RC3
Data sets
Water vapor measurements from Pico-Light H2O Mélanie Ghysels-Dubois https://doi.org/10.25326/M78G-AD37
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 308 | 100 | 61 | 469 | 65 | 60 |
- HTML: 308
- PDF: 100
- XML: 61
- Total: 469
- BibTeX: 65
- EndNote: 60
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The manuscript presents data from a balloon-borne campaign that integrated hygrometers, particle-measurement instruments, ozone sondes, and FIRMOS-B radiance observations. The intercomparison of the three hygrometers, together with a discussion of their respective limitations, is presented clearly. The measurements also identify distinct cloud and aerosol layers. Furthermore, the quality of the FIRMOS-B spectra is assessed, and preliminary retrievals of temperature and water vapour are presented.
However, the manuscript has several major shortcomings. The abstract and introduction lack clarity and are not adequately aligned with the substantive content of the manuscript. The abstract states that the work presented is useful for preparation activities of the ESA FORUM mission; however, the nature of these activities and the specific contribution of the present study are not explained. The final two sentences return to this objective but defer the substantive analysis to a subsequent publication: The assessment of far infrared spectral radiances measured by the FIRMOS-B instrument using the complementary information brought by in situ hygrometers and particle counters will allow to evaluate the quality of the retrievals and to raise questions on the influence of aerosols and optically thin cirrus on the far-infrared retrievals of FORUM-like observations. Such a deeper analysis will be the subject of a forthcoming paper. I assume the authors aim to answer questions rather than merely raise them. Moreover, the introduction neither develops the relevance of these preparatory activities nor situates the study adequately within the existing body of literature. Instead, it focuses predominantly on trends in water vapour without establishing how this topic relates to the campaign data presented in the manuscript. Particular emphasis is placed on the radiative impact of water vapour in the Tropical Tropopause Layer, although the campaign was conducted in Kiruna; the relevance of this discussion to the present study is therefore unclear. In addition, the introduction does not address aerosols, despite their prominent treatment in the results and conclusions, and gives comparatively limited attention to cirrus clouds.
I recommend substantial revision of the abstract, introduction, and conclusion to ensure that these sections accurately reflect the content and findings presented in the results section.
Additional major comments:
1) It is difficult to determine which instruments were deployed on each platform and the rationale underlying the respective deployment choices. A concise overview of the different platforms, provided before the individual instrument descriptions, would offer essential guidance to the reader.
2) The manuscript lacks a description of the FIRMOS retrieval methodology, including an explanation of how the water-vapour and temperature profiles are derived. Section 4.4 refers to a preliminary retrieval analysis; the methodology underpinning this analysis should be described elsewhere in the manuscript, preferably in the instrument section.