Review of “Observations of the downwelling far-infrared atmospheric emission at the Zugspitze observatory” by Palchetti et al.

A field campaign to measure the downwelling far-IR spectrum at the Zugspitze observations and ancillary information is presented in the manuscript. It is a timely campaign for the community given the ongoing far-IR mission implementations in both U.S. and Europe. High-quality far-IR ground measurements with sufficient ancillary information are extremely scarce, which hinders various efforts related to the satellite mission implementation. Because of this, the data described here will be precious to the community.

The manuscript is well organized and well written. I only have a few minor comments.

1. In Table 1 or the text, it would be useful to provide more instrument specification information—for example, the size of the field of view (FOV) for the FIRMOS and E-AERI. The NeSR for FIRMOS is presented in Figure 6, but no info about NeSR of E-AERI. Knowing the NeSR for both instruments will be of help to interpret Figure 7. Also, if the NeDT (Noise equivalent brightness temperature) can be given for the convenience of readers, it will be useful too for cross-literature comparisons.

2. The difference presented in Figure 7 is interesting. There are definitely some spectral features in the difference, which might suggest the frequencies of the two spectra are not perfectly aligned. FIRMOS and E-AERI are two FTS but with different spectral resolutions. It is well known that the instrument FOV can affect the spectral radiance measured by FTS (the off-axis contributions), which makes the comparison of spectra from different instruments challenging. A non-symmetric varying rectangular window can model this FOV effect on the frequency scale (see Bell 1972, Iacono & Clough et al., 1996). For FIRMOS and E-AERI, depending on their FOV size, the frequency scale shift is likely slightly different. But such a small shift in frequency can contribute to the not-small difference in spectral radiances shown in Figure 7. I am wondering whether this has been taken into account in the comparison.

References:

Bell, R.J., Introductory Fourier Transform Spectroscopy, 329 pp., Academic, San Diego, California, 1972. 

Iacono, M. J., and S. A. Clough (1996), Application of infrared interferometer spectrometer clear sky spectral radiance to investigations of climate variability, J. Geophys. Res., 101(D23), 29,439– 29,460.

3. Some acronyms in Table 1, such as PTH, PTHW, and SSA, can be written in full names instead of abbreviations. Though they are all defined in the text, it will save readers’ time if full names are given in the table.

This manuscript gives a nice overview of this Zugspitze deployment, its associated suite of measurements, and potential applications. It gives a good flavor of the clear sky measurements, showing a number of relevant figures and describing, although briefly, comparisons between the FIR instrument and the well-known E-AERI instrument. Given that this paper is meant to provide a review and roadmap for other scientists who might be interested in using this dataset, this brevity is fine and, even, expected.

The paper provides less extensive information with respect to how the dataset “can be used to constrain radiative properties of… cirrus ice particles, and snow/ice emissivity over almost all the infrared emission, including the under-explored FIR spectral range.” Figure 5 and the related description in the text are passable for the snow/ice emissivity, although showing an observed spectrum would be even better. However, the level of detail with respect to cloudy observations should really be increased. I’ll leave it to the authors to determine what other information should be included, but a figure showing contemporaneous measurements of FIRMOS and E-AERI would be helpful and welcome.

Minor corrections:

line 3 – Remove first comma.

6 - Remove comma.

7 – Remove both commas.

15 – “in the FIR that are not …”

20 – Since this first sentence is more of a definition, it probably should be combined with the next sentence, which has more content. i.e. “The Far-InfraRed (FIR), defined here as the longest wavelength region of the infrared spectrum covering the wavenumber range from 667 cm−1 (or equivalently 15 μm wavelength) down to 100 cm−1 (100 μm), contains more than 50% of the energy emitted by the Earth toward the space. This region …”

24 – Since “FIR” is already mentioned in this sentence, perhaps change the second “FIR” tp “gaseous”.

26 – “allow a better understanding of”

33 – “in preparation for”

40 – “wintertime”

41-42 – Perhaps change to “to the lowest wavenumber with potential semi-transparency, 250 cm-1. The sounding of the remainder of the FIR, below 250 cm-1, is …”.

50 – Remove comma.

75 – The values given for the shortwave end of the E-AERI measurements in wavenumber and wavelength are inconsistent.  1800 cm-1 is correct for the first channel of this instrument, while 3.3 microns is correct for the second channel. Probably the discussion here should explain this, and specify (if correct) that only the channel 1 radiances were used.

78 – Remove second comma.

114 – “100 km north of the Summit”.  “north” and “southeast” should not be capitalized.

124 – “allowed operations”

163-164 – “which causes a not-compensated phase error in FIRMOS measurements.”

170 – “form” should be “from”.

165-171 – I think this is implying that the E-AERI retrieval also used the 400-600 cm-1 region.  Please make this more clear.

Figure 5 – Please write  the meaning of “SSA” here since in atmospheric science it usually means something other than its meaning here.

This paper describes a dataset of infrared downwelling radiation measured with two FTS at the Zugspitze Observatory in Winter 2018 / 2019 together with further measurements of atmospheric parameters. It is a valuable dataset, covering a large spectral range from 100 to 1800 cm-1. Instruments and field campaign are in general well described, and the new FIRMOS instrument is successfully validated against the established E-AERI instrument.

With the focus on the calibrated spectra of infrared downwelling radiation, there are, however, some issues that should be clarified and discussed in more detail (see below).

I had some trobules when trying to work with the dataset. It is very cumbersome to download the complete dataset via the web page that was given for the review process; I am not sure if this is better via the ESA website. But I would stronlgy recommend to re-structure the data or enable an FTP access such that one does not have to click on each and every file for download (maybe there is a better way to get the data, but I did not find it). 

Specific comments:

Line 76: The authors give a spectral resolution of 0.3 cm-1 for FIRMOS and 0.5 cm-1 for E-AERI. No information is given on apodization and the definition of resolution. Is it defined as the FWHM of a delta line or something else? In the readme.aeri file it is said that the OPDmax is about 1.04 cm, and the instrumental line shape is an ideal sinc. I have not found the corresponding information for FIRMOS. Please indicate the maximum optical path difference and the applied apodization (if any) for both instruments in the text and comment on how the spectral resolution is defined.

Line 145: I am puzzled about the strong spectral structure in the calculated NESR. From all I know about FTS, I would expect a rather smooth curve, governed by the shape of the broadband instrument gain function. I am not fully convinced that the calculated NESR is in good agreement with the standard deviation of subsequent sky measurements. I tried to reproduce the comparison by taking 8 subsequent clear sky spectra, calculating the standard deviation and calculating the mean of the NESR given in the data. When comparing this NESR to the standard deviation, I find that the stadard deviation is generally smaller (as it can also be seen in Fig. 6 above 700 cm-1) and that the sharp lines below 350 cm-1 are way smaller in the standard deviation than in the NESR (which cannot be seen in Fig. 6, because the standard deviation is hidden behind the NESR in this range). Please explain in more detail why the NESR has these strong spectral features (way stronger than in the standard deviation) and tends to be larger than the standard deviation of subsequent measurements at high wavenumbers.

Line 154: It is stated that both instruments have a spectral calibration error of about 50 ppm. Since it is obviously possible to determine this error and to correct for it, why is this not done for the complete dataset but only for the comparison? A good spectral calibration would enhance the quality of the data.

Line 158: This is again in the context about resolution and apodization: Here it is said that the spectra are brought to the same spectral resolution of 0.5 cm-1 by applying a Norton-Beer strong apodization. If you apply a Norton-Beer strong apodization to an interferogram with an OPDmax of 1.04 cm (in the case of E-AERI), the resolution in the sense of the FWHM is clearly larger than 0.5 cm-1.

Line 160: The authors state that the FIRMOS and E-AERI measurements are in good agreement within the average CalErr estimate. In the dataset I found a note on an a-posteriori bias correction of the FIRMOS data, where the bias was determined from the difference between the FIRMOS and the E-AERI data. I think it is very important to make clear that the FIRMOS data has already been corrected using E-AERI data, because with this correction, the data is not independent anymore. Therefore this bias correction should be presented in the paper and it should be discussed how this bias shall be handled in future measurements when there is no E-AERI next to FIRMOS for comparison. Is this bias expected to be constant in time and under all conditions?

Line 163: Part of the difference between FIRMOS and E-AERI is explained by a non-compensated phase error in the FIRMOS data. Are there plans to compensate for this phase error in the future? If not, why not? If it will not be compensated, this error should be added to the estimated systematic error (CalErr).

Figure 2, lower panel: It would be interesting to see the residual in the overlapping range.

Figure 5: The center picture is not very informative without additional explanations. The right figure is very hard to read. I suggest to use different colors and/or symbols for the different measurements together with a legend indicating the measurement dates.

Table 1: Add dates of measurements for DUFISS

Snow Data: I am missing a readme file with the relevant information (including a contact person), instead I found some notes of a progress meeting which are only little helpful for understanding the data. In the data I did not find the error bars shown in figure 5, and it is not clear to me how to use the calibs.csv and the mV_g.csv files. In the file SSA_rho.csv I found the data points shown in the figure, but for 18-1751 it says (nan,nan), while I find some values in the figure. Please provide some more information along with the data.

Technical corrections:

line 121: remove the point after "properties"

line 147: blackbodies -> blackbody

line 150: the procedure

line 157: within 10 minutes

Figure 6: acquired on (remove "at")

Table 1: add a comma between H2O and O3 in the line of dedicated RS