This manuscript presents a much-needed investigation of Pandora HCHO products, including both direct-sun and sky-scan retrievals. Given there are limited studies evaluating these data products, particularly in Southeast Asia, the work addresses an important need. The authors analyze observations from five Pandora instruments across this region over several years, comparing variability between direct-sun and sky-scan measurements, as well as against satellite observations from OMI. While the study has the potential to make a valuable contribution, substantial revisions are necessary to address concerns related to the methodology and the interpretation of the results.

line 24-26: Sky-scan does not report total vertical columns. Only the tropospheric columns. 

line 81-87: I'm not sure this is true that previous studies don't differentiate between sky-scan and Direct sun modes for HCHO. I'd say it is more that there have been very few studies that look into DS HCHO at all because of known issues. Please reword this section. Same thing on line 121. 

line 113: Weird that you mention TEMPO and Sentinel-4 but not GEMS for a southeast Asia study? 

Table 1: State Altitude above sea level? 

line 142: sky-scan does not report the same product as DS. Sky-scan reports 'Tropospheric' column (usually 3-4 km) while DS reports total column. 

line 155: I suggest you reword this sentence because Pandora data quality flags are already confusing. All pandoras belong to the official pandora global network, but this does nothing to guarantee the quality of the data. 

line173-175: You don't need to include the file versions in the results. The methods is sufficient. 

Table 2: move to section 2.1 

Table 2: While explaining the data quality is necessary for Pandora discussions, I think there is more relevant information that would explain the Pandoras better. Because basically all data is unassured it doesn't do us much good to focus on that. Instead, I would rather see the 'high' 'medium' and 'low' flags that are also included in the L2 files. 

The DS HCHO is all unassured because the PGN does not have an official method of assuring that product (I'm not sure what is going on with Singapore, but if you have not I suggest reaching out to the operator to make sure the assured values are real).

Once the backlog of manually assuring the data is complete, much of the Sky-scan data should be fine, however this table does not show that. I am surprised that there is no assured data at all for any of these sites. 

Are these data only for the OMI overpass time?  

Line 188: Reword 

Line 189-197: No need to type out everything that is already in Table 2. 

Line 190: I think you should rethink your exclusion criteria. There is no official recommendation for excluding bad quality data, however several studies suggest removing data based on uncertainty (Rawat et al 2024). 

Table3: Don't need version numbers in description. 

Figure 2: What are the colors of the bars? If nothing, make the bars the same color. 

Figures 2 and 3: I suggest changing these to a normalized distribution plot. Because some Pandoras have more data than others we would be able to see the differences better. 

If the message is to compare DS to Sky-scan I also suggest including both results on the same subplot. For example a histogram of the one monitor's DS values in one color and the Sky-scan values in another color on top of that histogram. As it is now, it is hard to compare. 

Figure 4: In this figure description you change your filtering methods to also remove data above 50e15. This needs to be consistent throughout the entire results and stated in methodology. What is the reasoning behind this number? Instead I suggest filtering based on uncertainty and that would most likely give a similar result. 

You reuse letters in labeling the panels in this figure (and Figs 2-3). Each panel needs a unique letter. I also suggest combining the hourly and daily figures together so we can easily see the differences. 

Table 4: Under "remarks" are the distances needed? I don't understand what that is referring to other than the OMI spatial averaging column. 

Why noontime if the OMI overpass is closer to 1pm? What are the time windows for "daytime". Pandora does not report data at nighttime. 

Figure 5: x tick times need to be more clear. State month. 

Reword the description. The different experiments E1-E9 are referring to OMI not the Pandoras, right? 

I think you are being a little hasty in determining which experiment is performing best. This needs more discussion in the text. In Table 5 there does not seem to be a clear winner, rather each Pandora monitor works best with a different comparison method. 

Figure 6: No need for different colors in this figure. 

Line 315-319: This figure needs to be discussed in greater detail. I think this could be an important figure to show the comparison between OMI and Pandora HCHO distributions. 

Figures 7and 8: combine into one showing E1 in one color, and E2/E8 in another for easier comparison. Are these the daily comparisons? Noontime? 

lines 320-325: More discussion on this figure as well. This seems to be the main point of the paper yet only a few sentences.  

Section 4.1: This discussion and figure 9 should still be under results. You are presenting new information. 

Figure 10: I don't think a and b are necessary. 

the colors and markers on this figure are difficult to read 

Instead of SZA versus HCHO column, try SZA versus uncertainty. 

General/Conclusions: 

Why not include GEMS in this analysis? The time range should be perfect and there are several figures showing the hourly Pandora columns.

Further development into understanding the data quality (and the data products) is necessary. For example, it is not made clear in the manuscript that DS is total column, while sky-scan is only the lower portion of the troposphere. The Pandoras are also pointing in different directions throughout the day in the DS mode. 

Several figures could be removed/combined. 

This paper is presented as a comparison between Pandora and OMI, however the methods used for the comparison require further justification.