Dear authors,

After carefully reading the review reports and your response and revision to the manuscript, I think the paper can be accepted for publication in AMT, but I would like you to make the following modifications. Congratulations on a nice and comprehensive piece of work.

In the abstract of the revised manuscript (L11-12) you write "OMI always misses the frequently much higher values of TCNO2 that occur after the OMI overpass time". This should be rephrased as OMI is designed such that -by definition- it cannot measure TCNO2 after its overpass.

The next sentence should be nuanced or clarified. You write "OMI retrieved TCNO2 are not suitable for air quality assessments as related to human health, especially in polluted urban areas", but this statement is too sweeping. Please write exactly what you can justify based on your study, which is that OMI cannot resolve NO2 pollution within a city, which makes it less suitable for urban surface NO2 assesments (it can still provide some useful information if there is nothing else!).

L22: NO2 emissions --> NOx emissions

L43-44: "First, the mid-day OMI observations do not see the large diurnal variation of TCNO2 that usually occur after the 13:30 overpass time". Please rephrase such to make clear that OMI cannot be expected to see diurnal variation in the first place (unless when combined with another instrument, or at high latitudes).

L177: "In addition to missing the TCNO2 diurnal variation": again this should be reformulated. Readers should not get the impression that OMI was supposed to capture the TCNO2 diurnal variation.

L219-220: "The PANDORA values suggest upward airflow from the nearby circumferential ring road and resort areas".This statement comes a bit out of the blue. How do PANDORA values suggest this, why is it important? Please clarify.

Regarding the comments of Rev#2 and my earlier comments: I still think that Fig. 7 and 8 should be improved. They are not very clear, and require a lot of effort from a reader to see where the OMI time sits in the 3-D landscape. As suggested earlier, there are better ways to show the OMI 'underestimate' in the diurnal pattern. If you feel you really want to stick to Figure 7 and 8, please do so, but I think it hurts the clarity of the manuscript.

Last but not least, the thing still "hanging" with this work is the diurnal cycle in NO2 observed by the PANDORA, which often shows strong increases or maxima in the afternoon. Based on knowledge of the diurnal cycle in emissions, chemistry, and BL development this is quite surprising, and does not sit well with me. It would mean that we are totally overlooking an important process in the air pollution models, or that emissions are very, very different from what we think we know. But of course it could also mean that the PANDORA measurements have a weakness or not yet understood sensitivity. In the afternoon, the Sun starts to set, aerosols have been building up, and radiative transfer may not be as easy as in the morning or early afternoon, and this might affect the PANDORA retrieved values in the late afternoon.The manuscript needs at least a short paragraph discussing strengths/weaknesses/unknowns of the PANDORA method, also in view of other publications addressing the issue of diurnal variation in NO2. One obvious question that comes to mind is whether PANDORA measurements in the late afternoon have been validated or evaluated themselves (against MAX-DOAS). The other one is the concern on whether PANDORA does not measure too high NO2 at Mauna Loa ("more than possible in the stratosphere") - a further indication that something may be amiss with late afternoon measurements.

Congratulations on your paper.