This review is for egusphere-2025-5261, titled, Evaluation of Pandora HCHO and NO₂ with Airborne In Situ Observations. In this analysis, the authors evaluate Pandora HCHO and NO2 columns using in situ data from the NASA DC-8 during the AEROMMA period and also compare to TEMPO data. This evaluation is more a case study analysis rather than a statistical evaluation of the Pandora retrievals. Overall, the authors have done a dense analysis during these cases but is recommended for publication after a number of minor revisions are addressed as discussed below.

The analysis discussion going back and forth between NO2 results and HCHO results is confusing at times.  Please work on addressing ways to make it clearer to the reader which product is being discussed, particularly in the California case (e.g., 3.1.2 is about NO2 but has a lot of discussion of HCHO).

The results do not get compared to the expected accuracy and precision of the Pandora and TEMPO.  Consider adding this perspective to the analysis.

Line 17-18:  This sentence only talks about validation of Pandora NO2 and ignores HCHO validation.  Neither product from Pandora HCHO has been largely validated.

Line 26: The MAX-DOAS isn’t sensitive to the viewing angle, but the comparison is due to the large spatial variance in NO2.  The sentence is worded to imply that it’s the technique but instead it’s the characteristic about the variability and the mismatch spatially between measurements.

Line 30-32: Are these stats relative to DC8 or TEMPO?

Line 43-44: The MAX-DOAS also uses differential absorption.  The sentences implies it does not.  

Line 45-46: The final sentence in this paragraph does not fit with the rest of the paragraph.

Line 58: describe what is meant by drift.  

Line 79: MMS needs a reference

Section 2.1: Please describe why more of the AEROMMA dataset was not used over Pandora sites (e.g., Whittier).  It is likely criteria desired in the vertical completeness of the profile but explain the criteria used for identification of these cases from the larger AEROMMA dataset.  Are there any partial column comparisons that can be done at other sites?

Line 89: add “DOAS” before ‘fitting’

Line 90: clarify that the sentence is about DS.

Line 101: typo ‘Blick’

Line 102-103: This data source is not the same as in the data availability section. 

Section 2.3: This section mentions TEMPO NO2 but then does not explain why TEMPO NO2 is not used in this analysis like TEMPO HCHO.  Please add why this is not included and justify why only HCHO data is considered.

Line 225: clarify that the 6.6e15 value is from in situ integrated data.

Table 2: Clarify in the Table that DS NO2 has a climatological stratospheric value subtracted.

Line 237: clarify that its MAX-DOAS if talking about the column but if talking about the profile, it doesn’t appear to be 50% different throughout the entire profile.

Line 249: add ‘than in situ’ after ‘lower’

Line 251: clarify that its 40% lower for the MAX-DOAS Pandora tropospheric column. 

Line 257 and the preceding lines: there is also an uncertainty in the climatological stratosphere that would be an important factor, especially at such low column values.

Line 261: the partial columns do not get anywhere near the tropopause, so this statement does not make sense.

Section 3.2: Is there overlapping data in these flights? The flight summaries from the NOAA AEROMMA website talk about doing the pattern in reverse on the way back.  If so, should these be separated? Otherwise, discuss why not or why only one pass is shown.   

Bottom of Page 12: State the maximum HCHO measured on NYC1 and NYC 2 to be comparative with NYC3 discussion.  Winds are not mentioned either which is relevant to the discussion of Figure 5.

Line 309: Is this not also evident in the DS HCHO too?

Paragraph starting on Line 321.  As the AEROMMA aircraft had two passes of sites each day, is there any work that can be done to evaluate the 50% growth vs. 125% growth between DS and MAX-DOAS?

Line 329-331: While there is more data remaining, there would still need to be validation between cloudy and clear sky scenes for MAX-DOAS. Qualitative analysis is still valid but would need accuracy validation for quantitative analysis.

Figure 5: It is clear there is something wrong with Manhattan DS retrievals.  Any indication what is going on there?

Line 365-366: why are the buildings of NYC mentioned in this paragraph if no vertical profiles were collected over NYC? The sentence seems out of place.

Line 371: are in situ columns not extrapolated to the ground for the column calculation? They should be, but the explanation of the difference implied that it is not. If they are, then please better explain what is meant by the difference being from larger partial columns in the lower BL.

Line 371: Also the statement in this sentence isn’t true for NYC2. Which is stated in the next paragraph, but it just makes the first sentence starting in Line 371 inaccurate.

Line 397-398: Why is the NewHavenCT profile on NYC3 binned differently and how?

Line 400: what are the primary sources of HCHO here? Add references that state this is a factor in the NYC area?

In the Pandora section, make it clear whether the tropospheric columns from the MAX-DOAS profiles are the same as the sum of the partial columns and if they’re derived separately.

Figure 5 and Line 427: Are these TEMPO data that are collocated over the Pandora sites or are they over the whole domain shown in Figure 7? If the latter, this is not a valid comparison as the majority is over water where on NYC4 you show that all the high levels are over land.  This is skewing the analysis. They should not be quantitively compared in this way.

Line 442: MAX-DOAS still needs validation under these partially cloudy scenes. But this discussion puts all the emphasis on the struggle of TEMPO or satellites in general with clouds. While there is data for evaluation of MAX-DOAS on this day from Figure 6, this appears to be after the clouds clear in the afternoon hours.

Figure 8:  Please clarify in the discussion that each point represents a TEMPO scan time (if this is the case).  However, it implies that TEMPO is averaged over the whole domain rather than at the Pandora sites.  As stated above, this is not a valid method for quantitative comparisons.  Data should only be considered over the Pandora sites and a spatial matching criteria should be stated.

Line 469 states a slope of 0.93 and the figure and the table do not even show this value. Same with the stated intercept.

Line 471: Why is there not a reason to believe the bias is present in Pandora and instead is stated as all on TEMPO? There is not enough evidence here to pinpoint which dataset is accurate, if any of them.