the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 11 Aug 2026
Evaluating vertically resolved Pandora formaldehyde retrievals using airborne and ground-based in situ measurements
Abstract. Pandora formaldehyde (HCHO) retrievals in the boundary layer provide continuous observations of HCHO vertical distribution, yet validation of these retrievals remains limited. This work evaluates Pandora retrievals of the vertical distribution of formaldehyde (HCHO) against aircraft in situ measurements collected at four U.S. East Coast sites during the summers of 2024 and 2025. Tropospheric columns and vertical profiles are consistent with aircraft-based observations within measurement uncertainty, with agreement that depends critically on the alignment between the Pandora viewing geometry and local horizontal concentration gradients. We further demonstrate a revised interpretation of vertical profile layer heights that improves agreement with aircraft profiles. These results support the use of Pandora HCHO observations for model evaluation and tropospheric chemistry applications.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Atmospheric Measurement Techniques.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(2311 KB) - Metadata XML
-
Supplement
(389 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'review of "Evaluating vertically 1 resolved Pandora formaldehyde retrievals using airborne and ground-based in situ measurements" by Pandey et al', Anonymous Referee #1, 01 Sep 2026
-
RC2: 'Comment on egusphere-2026-3479', Anonymous Referee #2, 08 Sep 2026
This article uses in-situ aircraft observations of formaldehyde (HCHO) to validate Pandora HCHO retrievals in the boundary layer and the troposphere. The study focuses on four U.S. East Coast sites during the Summers of 2024 and 2025. The period of study is dictated by the availability of correlative data obtained during the Student Airborne Research Program (SARP) campaigns in 2024 and 2025 and the Associating Local Emissions of trace Gases with Regional Observations from Satellites (ALEGROS) campaign in 2025. These aircraft campaigns carried on spiral flight patterns within the Pandora instruments viewing sectors, providing in-situ vertical distributions of HCHO in the lower troposphere. The aircraft observations allow for the characterization of the Pandora retrieval performance.
The study finds that Pandora tropospheric columns and vertical profiles are consistent with the aircraft observations to the extent expected given Pandora retrieval uncertainties and the inherent difficulty in matching observations resulting in representativeness issues. However, this statement is not fully supported by the information provided in the manuscript.
For years the community has been using Pandora and MAX-DOAS observations to evaluate the performance of HCHO satellite retrievals and model simulations. For years, the question of Pandora and MAX-DOAS performance has been raised. Therefore, this study together with the recently published analysis by Sebol et al., 2025 are welcome additions to the peer review literature addressing a fundamental question given the use of Pandora and MAX-DOAS networks in the validation of satellite retrievals of HCHO, particularly from geostationary platforms.
Given the considerations made above I recommend the article for publication. However, it would be beneficial if the authors could address the following questions. The comments below suggest adding further details to the manuscript to provide better context in the introduction, clarify the methodology and expand the discussions to provide quantitative conclusions.
Major Comments
- The introduction, and the paper in general, lack details regarding previous work. An enhanced discussion of the state of the field and the inclusion of associated references would provide readers with better context for the results presented later on.
For example, while discussing validation studies of satellites HCHO retrievals should be more specific. For example use https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2026JD046497 and https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2025JD044788?af=R for TEMPO, https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2024EA003894 and https://www.sciencedirect.com/science/article/pii/S0048969725008253 for GEMS, https://doi.org/10.1021/acsestair.5c00223 and https://mpc-vdaf.tropomi.eu/formaldehyde?start=1 for TROPOMI. The first and last articles provide an insight on the differences associated with FTIR or UV (MAX-DOAS, Pandora) retrievals and the need for careful interpretation.
This is only one of the aspects in which the introduction could be expanded to provide more context. Please revise the section to also better convey the importance of HCHO measurements to constraint emissions, characterize the oxidative state of the atmosphere, air quality applications, and atmospheric chemistry by providing further references. - In the Methods section (subsection 2.1) the benefit of measuring angles 75 to 60 to gain sensitivity to lower troposphere columns is explicitly mentioned. It would be great to expand the discussion to explain what other zenith angles provide and include a reference to the study supporting such a statement.
Likewise, subsection 2.1.1 lacks some detail (not clearly provided also in the Appendix) and could use a more detailed description of equations 1, 2 and 3. For example, how is VCDO2-O2 obtained? What about nsurfaceO2-O2?
Subsection 2.1.2, what is the relationship between the filtering criteria (fitting residual RMS, relative uncertainty of tropospheric column and surface mixing ration) and cloud contamination. Are there proven relationships between the presence of clouds and these three quantities? Is there a reference that justifies the selection of thresholds. - In the Methods section (subsection 2.2) it would be helpful to further discuss the quality control and if there was need for further calibration of the CAFE and Aeris instruments during the campaigns, particularly the Aeris instrument given the information provided by Mouat et al., 2024 (https://amt.copernicus.org/articles/17/1979/2024/).
- Results and Discussion. In general the discussion can be expanded to extract quantitative conclusions using the data already shown. For example, where there are no SW profiles at RCC? Why are they not shown in Figure 3? Would it be possible to provide a linear fit equation in figure 4 with its slope and intercept when considering all spirals and leaving out those for which the lowest altitude was > 500 m. In figure 5, if available, including information about the prevailing winds will be interesting for interpretation and to confirm the origin of the air mass sampled in-situ.
As mentioned earlier, figure 6 and its associated discussion would benefit from more quantitative analysis, showing the differences for each later between the in-situ and Pandora retrievals and quantifying the biases for the operational and modified Pandora retrievals.
Adding further to the “Results and Discussions” section, providing quantitative overall results and discussing their uncertainties will be extremely helpful for model or satellite users when comparing with Pandora observations. It is possible that the dataset used here is too small to reach broad conclusions but that could also be explained. - Please modify the conclusions to provide some quantitative results reflecting the analysis suggested in point 4. Are there plans to use the revised PZA-to-height mapping formulation operationally? Because of its interest for users of the Pandora retrievals, can the PZA-to-height mapping be recalculated with the information contained in the Pandora operational files. Including this discussion in the paper (either in the conclusions) or as part of the Appendix will be helpful.
Finally, it would be good if the results of this analysis could be placed in the context of the work recently published by Sebol et al.
Minor Comments- Abstract (line 25): may be interesting to mention satellite validation in the uses of Pandora HCHO observations.
- Introduction (line 38): May be better to use Zoogman et al., 2017 to refer to TEMPO https://www.sciencedirect.com/science/article/abs/pii/S0022407316300863
- Methods (line 84): Please revise the grammar of “Scans at higher elevations, namely 75° and 60°, constrain the column abundance in the lower tropospheric.”
- Methods (line 99): Please clarify, is the “NASA real time algorithm” the same that is described in Cede (2024)?
- Methods (line 174): Why was the time coincidence criterion relaxed to +-30 minutes for the Essex site?
- Results and discussion (line 264): “One profile comparison in Fig. 6 (panel f) shows a marked difference between even the modified Pandora profile and the in situ measurements.” Revise grammar, what is meant by “...between even the modified Pandora…”
Citation: https://doi.org/10.5194/egusphere-2026-3479-RC2 - The introduction, and the paper in general, lack details regarding previous work. An enhanced discussion of the state of the field and the inclusion of associated references would provide readers with better context for the results presented later on.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 145 | 70 | 18 | 233 | 27 | 15 | 14 |
- HTML: 145
- PDF: 70
- XML: 18
- Total: 233
- Supplement: 27
- BibTeX: 15
- EndNote: 14
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The paper focus on evaluating Pandora HCHO MAX-DOAS measurements against reference aircraft in-situ measurements from 2 aircraft campaigns over four sites in the Easetern US, and propose a revised approach for the vertical layer mapping part of the operational NASA near-real time processing. This part is key to retrieve profiles in the low tropopshere from the Pandora slant columns measurements, and the study illustrates how the new approach show more consistent results with the aircraft data.
The topic is relevant for the scientific community, with the fast spread of the Pandora instruments and their use for GEO and LEO satellites validaton, but with only a limited validation of the Pandora data themselves.
By its nature (comparison to aircraft campaign measurements), the paper presents results from a few sites only (4 sites, few days in summer time, 30 profiles), and is in my view a bit too short/quick in the conclusions (recommend to use the new vertical layer mapping for the whole PGN processing) based on this limited dataset. I agree that for the data presented here the comparisons seems better with the new approach, but how representative are they for the rest of the PGN network? Moreover the results are mostly presented on a qualitative basis, with figures for a few cases only, and the paper lacks, in my view, a more quantitative assessment of the Pandora profiles quality and of the improvements that the proposed method bring.
The subject is interesting, novel and well presented, and is in scope of the journal.
I suggest publication after minor revision including the addition of a few quantitative metrics (slope and intercept of regression plots, quantification of agreement and improvement per profile layer, ... see below) to present the results, and some discussion of the limitations/representativeness of the shown results (only 4 sites, only summer conditions, ...). I suggest to add some references to recent HCHO Pandora validation/exploration/comparisons efforts (ie TEMPO validation by Rawarat et al. 2026, Ortega et al. 2026, TROPOMI validation by Park et al. 2026, Pandora sun and sky comparisons by Chi et al. 2025, see below) and comment them a bit. Ideally, it would be nice to see the impact of the new vertical layer mapping to the data of the parallel Sebol et al., 2026 study, or at least mention it as a perspective to confirm the results shown here before recommendation of applying the new method to the whole network/operational algorithm.
technical+minor remarks
-------------------------------------
- line 37: "that are widely used for validation of satelltie products" -> I would give more references here of the used of Pandora for validation of HCHO satellite products, to justify the use of "widely"... some examples: Park et al. 2026, Rawarat et al. 2026, Ortega et al. 2026, Chi et al. 2025
- line 47: "MAX-DOAS retrievals employ the optimal estimation approach" --> parametrezide approaches are also quite common for MAX-DOAS profile retrievals. Please mention them.
- line 56: update the Sebol et al. reference (2025 -> 2026)
- line 84: please refer to the PGN Level 2 version. Either here, either when introducing the modified fitting window in lines 92-95. How much this change of the fitting window change the DSCD themselves? Fig S1 show the reduction of the DSCD error, but is this also affecting the DSCD (and thus the HCHO VCD and profiles and surface values)? Please add some quantitative comment on how much this fitting choice change the values compared to the values obtained with the operational algorithm.
- line 125: refer to appendix A when mentioning the scaling of the partial columns to the tropospheric vertical column.
- line 128: "MAX-DOAS sensitivity kernels are strongly asymmetric and bottom-weighted, with most
sensitivity concentrated near the surface and a long, weak tail extending to higher altitudes." --> Please provide some references (e.g. to figures in optimal estimation retrieval papers ) for this statement.
- lines 156-160: why not using the L2_qa values to filter the data? How much would the results presented here change if the "classical filtering method" would be used? Please comment and refer to recent papers showing the limitations of the operational automatic flagging algorithm (Rawat et al. 2025; 2026, Arul et al., 2026). Maybe also add the column number for the variables you use.
In a typical L2_rfuh5p1-8 file we have:
'Column 10: rms of unweighted fitting residuals, -9=fitting not successful '
'Column 11: Normalized rms of fitting residuals weighted with independent uncertainty, -9=fitting not successful or no uncertainty given '
'Column 12: Expected rms of unweighted fitting residuals based on independent uncertainty, -9=fitting not successful or no uncertainty given '
'Column 13: Expected normalized rms of weighted fitting residuals based on independent uncertainty, -9=fitting not successful or no uncertainty given '
'Column 42: L2 data quality flag for formaldehyde, 0=assured high quality, 1=assured medium quality, 2=assured low quality, 10=not-assured high quality, 11=not-assured medium quality, 12=not-assured low quality, 20=unusable high quality, 21=unusable medium quality, 22=unusable low quality '
'Column 46: Independent uncertainty of formaldehyde surface concentration [mol/m3], -6=no surface concentration was retrieved since the maximum viewing zenith angle was below 87deg, -7=uncertainty could not be retrieved since slant column uncertainties were missing '
'Column 50: Independent uncertainty of formaldehyde tropospheric vertical column amount [moles per square meter], -4=tropospheric column was estimated from direct sun measurements using effective temperature, -5=tropospheric column was estimated from direct sun measurements using stratospheric climatology, -7=uncertainty could not be retrieved since slant column uncertainties were missing '
- line 171: please give a reference or a website, ... for the "Aeris Ultra" in-situ sampler. What is its uncertainty?
- line 182: "aircraft-measured pressure and temperature were used to convert Pandora partial columns (mol m-2) to volumetric mixing ratios (ppb)". Please refer to Sebol et al., 2026 that made the test of aircraft-measured vs climatological values showing that "The two methods vary by less than 5% (Figure. S3)"
- line 200, Fig 3: "These observations highlight the importance of sampling geometry when comparing Pandora observations with in situ measurements" --> this is very interesting, as often we assume HCHO to be relatively spatially homogeneous (ie compared to NO2). If you have wind informations from the aircraft, it would be nice to assess if this case is a specific high wind condition or if we expect a spatially heterogeneous situation due to the Pandora site configuration (city vs more vegetation, ...), ... [same comment for Fig. 5]
- Figure 4: To increase the quantitative aspect of the study, please add regression analysis statistics, such as slope and intercept, in addition to person correlation coefficient.
- line 223: "collocated Aeris measurements for three days each during the summers of 2024 and 2025" --> how long is the parallel measurementsment of the Aeris in-situ sampler? only the days shown? only during the flights or for a longer period? If yes, this would be nice to use/show...
- line 225 and 240: "As with tropospheric column comparisons, spatial variability influences agreement between the two datasets." --> this is to be expected, as the near-surface concentration is retrieved from the low elevation angles, which have long horizontal sensitivities (in opposite directions). Maybe you can provide the typical horizontal extent values, as provided in the Pandora files, to give the reader a bit more of context on how different are the in-situ point measurement and the integrated along the line of sight near-surface Pandora measurement.
- Figure 5: add in the caption that this is the RRS site. Please also confirm if the error bars on the Pandora are the errors as provided in the Pandora files.
- line 256: "the operational Pandora layer mapping" --> did I understand well that the black is not the operational PGN data (but using the operational approach), because of the improved DOAS fitting? How different would the operational data be in Figure 6?
- line 261: "This produces a vertical shape that bulges out in the first few 100 m above the surface and drops to very low values above ~1 km." --> also refer to Sebol et al. study, that also showed similar shape problems with other aircraft measurements.
- line 277: "The consistency between retrieved and observed vertical shapes is broadly reproduced across the full set of campaign sites." --> please be more quantitative than "broadly". You could provide an estimate of the partial columns for some of the different vertical layers? at least for Figure 7...
Same comment for line 283: "the degree of agreement"...
- line 296: "Pandora tropospheric columns are consistent with aircraft-integrated columns within measurement uncertainty" --> be more quantitative. I have not seen comparisons including the measurement uncertainty... reword or add more in the manuscript.
- line 305: "We recommend adopting the revised height mapping in operational PGN retrievals, to avoid systematic concentration of the HCHO column into the lowest few retrieval layers." --> although the cases shown here seems to be convincing, how representative are they of the other Pandora cases? this is only 4 sites and only summer measurements, over 30 flight spiral... I would first recommend a larger scale analysis of the profiles, with testing of this new approach, before recommending a full switch... or maybe you already know other examples (ie the Sebol paper, maybe some of the GEMS validation flights over Korea, ...), but here it is a bit quick, in my view.
- line 310: "and other GEO instruments" --> also for LEO validation!
- line 349: maybe add "The tropospheric column "q" is estimated..." --> I don't think it was introduced before...
references:
--------------
Ortega, I., Hannigan, J. W., Edwards, D., Stremme, W., Grutter, M., Cadena-Caicedo, A., et al. (2026). Evaluating TEMPO formaldehyde retrievals with co-located ground-based FTIR and Pandora observations. Journal of Geophysical Research: Atmospheres, 131, e2026JD046497. https://doi.org/10.1029/2026JD046497
Rawat, P., Crawford, J. H., Travis, K. R., Judd, L. M., Demetillo, M. A. G., Valin, L. C., et al. (2025). Maximizing the scientific application of Pandora column observations of HCHO and NO2. Atmospheric Measurement Techniques, 18(13), 2899–2917. https://doi.org/10.5194/amt-18-2899-2025
Rawat, P., Travis, K. R., Henderson, B., Crawford, J. H., Judd, L. M., Demetillo, M. A. G., et al. (2026). Spatiotemporal assessment of the TEMPO formaldehyde column retrieval using the Pandonia global network. Journal of Geophysical Research: Atmospheres, 131, e2025JD044788. https://doi.org/10.1029/2025JD044788
Arul, S. A. L. D., Chang, J. H.-W., Wong, Y. J., Ooi, M. C.-G., Liew, J., Chee, F. P., Dayou, J., Sentian, J., Aryastana, P., and Lin, N.-H.: Direct-sun versus Sky-Scan Pandora Formaldehyde Retrievals: Implications for OMI Validation in Tropical Southeast Asia, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2026-716, 2026.
Chi, N. D. T., Tanimoto, H., Inomata, S., Fujinawa, T., Müller, A., Sugita, T., et al. (2025). Characterization of Pandora sky-scan modes from the Japan Pandora Network and evaluation of their use in validating satellite NO2 and HCHO retrievals. ACS ES&T Air, 2(12), 2872–2888. https://doi.org/10.1021/acsestair.5c00223
Sebol, A., Wolfe, G. M., Canty, T., St. Clair, J., Delaria, E., Kaiser, J., Desai, N., Rollins, A., Waxman, E., Zuraski, K., Place, B., Pandey, A., Singh, A., Ring, A., Gatebe, C., and Dean-Day, J.: Evaluation of Pandora HCHO and NO2 with airborne in situ observations, Atmos. Meas. Tech., 19, 5539–5552, https://doi.org/10.5194/amt-19-5539-2026, 2026.
Jong-Uk Park, Subin Lim, Thomas F. Hanisco, Nader Abuhassan, Bryan K. Place, Apoorva Pandey, Alexander Cede, Martin Tiefengraber, Manuel Gebetsberger, Jinsoo Park, Jinsoo Choi, James H. Crawford, Chang-Keun Song, Sang-Woo Kim, Global analysis of nitrogen dioxide and formaldehyde column densities from the Pandora global network: Variability and implications for satellite validation, Remote Sensing of Environment, Volume 335,
2026, 115249, ISSN 0034-4257, https://doi.org/10.1016/j.rse.2026.115249.