Please refer to the comments in the supplement PDF.

Liao et al., utilize a four-season deployment of Atom aircraft observations to validate three HCHO retrieval products. They demonstrate that these HCHO products generally capture the spatial and seasonal distribution of HCHO in the remote ocean-atmosphere albeit with a low bias. An important result of this study is that the biases in slant column corrections have larger impacts on retrieval than AMFs. The paper is well-organized and includes technical details that fit well into the scope of AMT. I hope the authors can address the following comments before the paper is accepted for publication in AMT.

Major comments:

The conclusion of this paper could be further strengthened. While the study effectively validates these HCHO retrievals and addresses differences in HCHO columns across latitudes and seasons, it would be valuable for the authors to provide practical advice to users of these products. For example, do the authors have any recommendations on which retrieval product is preferable? Would averaging across multiple products yield more accurate results than using a single product? Alternatively, should the spread among the three products be treated as an indicator of uncertainty in HCHO retrieval?

Other comments:

1. The OMI satellite overpass time is 1:30 pm local time while the Atom observations were conducted throughout the day. How do you account for the time difference when comparing HCHO retrievals to Atom observations?
2. Line 121: the ascents and descents of aircraft measurement cover 200-450 km in horizontal distance, which is larger than the pixel size of satellite retrievals. Also, the aircraft provides in-situ measurements while the satellite measures pixel by pixel. How do you account for the differences in the spatial scales of these two observations?
3. It is unclear how you treat cloudy conditions when mapping satellite retrievals to ATOM observations. Do you only select satellite/ATOM observation under clear sky conditions?
4. Figure 2: why is there an enhancement of the HCHO column at ~ -60 latitude bins in the OMI BIRA retrieval products?
5. Line 313: since negative bias is more pronounced at higher latitudes, does it suggest that the latitude-dependent background correction is insufficient?
6. Line 367-368: what does “variability” refer to here? If it refers to uncertainties, a factor of 10 seems too large. If it refers to the full range of corrected slant columns, I don’t understand why this implies that uncertainties in AMF are a minor contributor to overall retrieval error