Precipitable water vapor retrievals using a ground infrared sky camera in subtropical South America by Elion Hack

This work analyzed IR imagery produced by the ASIVA sky camera to measure the downwelling radiance at 10-12 µm, L_λ.  The method for retrieving the atmospheric PWV in this study used the relative distribution of humidity profile climatology of the sampling location. The results can be useful to applications seeking to study the role of spatial-temporal transformations of water vapor in the atmosphere, especially in time-sensitive processes such as the initiation of convection. In that respect the present study has a high potential for publication after the incorporation of the comments/suggestions as given below:

1. You have used Radiosonde data for retrieving the atmospheric PWV. But the Radiosonde and sunphotometer data are also associated with errors. Explain the possible sources of Radiosonde and sunphotometer errors in your analysis.
2. How were these thresholds determined to identify clear or partially cloudy pixels? You should include some explanation of these values.
3. There are three questions, which might be interesting for the readers if you could answer them in the conclusions:

i) Overall, do you consider the quality of the Precipitable water vapor retrievals using a ground infrared sky camera to be "good" or "satisfying"? Is the quality good enough to provide benefits for users as well as a repository of data?

ii) Could the Precipitable water vapor possibly be improved in future work?

iii) How does the quality of Precipitable water vapor retrievals using a ground infrared sky camera compare to earlier studies? How do the validation results of this study compare to earlier work?

Review on Precipitable water vapor retrievals using a ground infrared sky camera in subtropical South America.

The paper address an important scientific question, as water vapor, an important trace gas, is known to be difficult to represent spatially and temporally with current instrumentation and this instrumentation could help in this regard. The idea of the paper is novel, taking from Mims et al. (2011) idea of using IR thermometer to point to the sky, but improving it with a camera that can give us spatial distribution information. The conclusions could be extended a bit (see my comment below), and I cannot say they are substantial but they are a first approximation to the problem of retrieving water vapor information with a ground infrared sky camera. The methods are clearly outlined and, in my opinion, the assumptions are valid. The results are just a start (one site, some periods of a few days, ...), but I think they are enough for a first approximation to the topic and determining the best approach for extended studies. Regarding the description of the experiments, I think that some things are missing but in general the description is good (see my specific comments below). Title, abstract, and bibliography are correct in my opinion, and the presentation is well structured and clear. The language is generally correct in my opinion (see technical issues for some small corrections). I therefore recommend minor changes in the paper before it is ready for publication.

 Specific comments

It is not clear how the synthetic profiles are obtained. It would be interesting also to indicate the equations of these profiles.

L178-179. It is claimed that averaged profiles are also used but then in lines 185-189 this is not mentioned.

L196-197. It is claimed that cloudy pixels are remove, but it is not stated how (by visual inspection, with a threshold value in L, ...).

In general, I am missing some table showing the statistics in the different methodologies/situations to compare them to each other easily. The conclusions could also benefit from this kind of information.

The flowchart in Figure 3 looks very clear to me, so I congratulate the authors for it.

 Technical issues

1. L.155 measured --> compared

2. L.158. I am not sure this is completely correct, there can be places in which winter is dry and with clouds. It also depends on what "dry" really means (less relative humidity? less specific humidity?).

3. L296-298. Maybe rewrite the sentence as: "Thus, provided that both the temporal trend and variability from the two series can be equivalent,...".