|The reviewed Revision 5 is significantly improved over Revision 3. Sections are better organized and balanced, and more background information is provided to contextualize the presentation. While no further rewrite of the paper is needed, I have several comments that I would like addressed prior to publication.|
The paper contains references to rows and columns throughout. While such terminology is appropriate when discussing detector effects, it is less so for issues such as radiometric response. Some figures, especially but not limited to Fig. 18, would be more informative if plotted versus wavelength.
The authors are still underplaying the importance of the BSDF calibration. The revised discussion makes clear there were pre-launch calibration problems that prevented a high-quality set of BSDF calibrations. But the instrument still has BSDF calibrations, as presented in Section 6.3. It is noteworthy that this section precedes sections dealing with relative radiance and irradiance, suggesting that these somehow do not affect the BSDF. There are three basic spectral frequency regimes of interest for the BSDF, listed with increasing importance to science products: wavelength-independent errors, broad wavelength dependence, and spectral structure. The authors present a relative BSDF uncertainty for each sensor channel, but do not say over what spectral frequency these values apply. The authors state the BSDFs are spectrally smooth, so the stated uncertainties clearly do not describe residual spectral structure. Fig. 18 suggests (it would benefit from improved quality) the BSDFs are not perfectly smooth, so the authors should provide an uncertainty component for the residual structure.
This section or the conclusions would benefit from some critical discussion of the results. By simply looking at Fig. 18 and with some knowledge of aluminum reflectance it is possible to see significant problems in the UV channel. Are the other channels better understood?
Section 6.6, entitled In-band straylight, contains a nice description of terminology but never tells us what in-band refers to. Is it all stray light that is not out-of-range? Or all stray light that is not out-of-band? The authors (in Section 6.7) seem to use these two terms interchangeably. In other words does in-band stray light ignore inter-band stray light, e.g. from VIS wavelengths into UV wavelengths?
Section 6.6 contains the statement, "The SLRF describes the relative straylight response of the system and is derived from on-ground calibration measurements." Which measurements were used and how were they used?
The discussion in Section 6.6 centers on the EWLS measurements, which evidently did not involve spectral cutoff filters. The omission of information about spectral stray light is glaring, regardless of how the requirements were defined. The numbers shown in Table 8 and in Fig. 25 are therefore rather misleading to the average reader. The authors state at the end of the section that laser PSFs form the basis for the processing correction, yet no quantitative assessment is provided. Even if the authors cannot estimate the effectiveness of the correction, they should report the total stray light correction at several wavelengths within each band.
Table 8 needs more explanation, either in the caption or in the text. What is the definition of stray light percent (what is the value in the denominator)? Same comment applies to Fig. 25.
Figure 17 appears to be mislabeled. Different lines should be columns instead of rows.
Section 6.6, paragraph 3. Tol et al. (2018) should be followed by a period or semi-colon.
Figure 23 caption. The fifth sentence is poorly formed.