Dear Authors,

Thank you for the revised manuscript.

I think most of the suggested comments have taken into account.

I have still one major issue that I think it has to be addressed in order for this paper to be published. For the reviewers point of view but also mine there is too much effort in order to compare the local noon values.

In principle to compare these values is useful but it is common sense what it is shown in your figure 5a: that local noon OMI values can never be lower than overpass values. This is because if there are clouds in the overpass time OMI assumes that they are the same clouds also at noon and if there are no clouds then it is assumed that there are no clouds in at noon and solar zenith angle affects this ratio. So in the real atmosphere there are chances that clouds exist at noon and not in overpass time (1) and vice versa (2). So if statistically cases 1 and 2 are equally probable over this huge dataset, what you show in these figures is more or less expected. And the seasonal variations of these rations too as in the summer time you expect less clouds so less cases 1 and 2. Changes on the above ratios when for some local reason there is more probability of cloudy conditions e.g. at noon will lead to different results, but this is not the scope of this paper I guess..

What I am saying is that the whole paper should be more focused (abstract, figures, conclusions) in the overpass data. It would be useful to highlight that there is a X% overall overestimation of OMI compared with ground for overpass spectra and a table or figure to show this overestimation ( for example as a median of the ratio OMI vs ground ) plus the standard deviations etc. So this is the main finding. Then for the noon measurements someone can say that what we expect is more scatter due to the OMI assumptions that have been discussed above.

So summarizing, to focus more on the overpass results and description and less on the local noon ones (that can be also presented). Temporal and spatial analysia and trends are also important as presented .

Minor

Line 172
there is also this new reference to have a look
Zempila, M.M., et al ., Validation of OMI erythemal doses with multi-sensor ground-based measurements in Thessaloniki, Greece, Atmospheric Environment, doi: 10.1016/j.atmosenv.2018.04.012, 2018

and line 172.
This one for the spatial and temporal issues:
S. Kazadzis, A. Bais, D. Balis, N. Kouremeti, M. Zempila, A. Arola, E. Giannakaki, A. Kazantzidis, V. Amiridis, Spatial and temporal UV irradiance and aerosol variability within the area of an OMI satellite pixel, Atm. Chem. and Phys., 9, 7273-7298, 2009

best regards
SK

I think the paper has been improved a lot and it is a very good contribution on solar UV satellite based validation against a large number of instruments/sites.

Congratulations for your work.

Just for discussing a bit more the overpass vs local noon aspect.
What is shown here is that overpass and local noon rations of OMI/GB are ~1.07 or ~+7%
as it is written in the abstract for both (OV and LN). That is fine.

In addition you show that the results of the LN/OV ratio. there OMI (1.18 vs 1.38) underestimates. As OMI takes into account only solar zenith angle change (LN vs OV) it means that statistically at local noon in all sites you have less clouds or less total column ozone compared with the overpass.

But statistically this is not shown in this 7% ratio for both discussed in the abstract.
Someone would expect that since OMI underestimates the local noon value the local noon OMI/GB ratio should be less than 7%.

I guess it is that more complex statistics involed. But since this is one of the main "messages" of the paper, i just had to mention it, if you want to include some comment on it.

best regards
SK