Review of the revised manuscript by Loyola et al.
I think that the authors did not adequately address, at least, two comments which are important for evaluation of the algorithm error budget.
1. The authors state that the GB model was tested against the full RGB OCRA model and they found “only a small mean bias of 0.07” in the retrieved cloud fractions. No figure that illustrates the cloud fraction differences is provided. The cloud product is intended for the use in trace gas retrievals. For this purpose, the low cloud fraction range is most important. The mean bias of 0.07 is not small for low cloud fractions; it is quite large. The authors should compare cloud fractions from RGB and GB in more details and provide corresponding illustration. For example, they should look at the dependence of the mean bias on the RGB cloud fraction. The GB model may appear not to be appropriate for TROPOMI. Moreover, the red channel (675-775 nm) is available in TROPOMI; that is why the implementation of the RGB model has no problem for TROPOMI.
2. The authors removed the unsupported statement that “the use of water or ice properties has a relatively small impact on the O2 A-band spectral region”. However, they ignored my suggestion to compare the simulated TOA radiances and provide an estimate of cloud pressure errors due to the use of a single scattering phase function for water clouds. The authors answered “Mie theory is not sufficient to describe the scattering from ice crystals”. This is true. However the radiative transfer simulations do not necessarily require the Mie theory. The simulations do need a scattering phase function. Scattering phase function for ice clouds are available, e.g. at http://www.ssec.wisc.edu/∼baum/Cirrus/IceCloudModels.html.
I recommend the manuscript for publication after the authors address these comments. |