This study examines actinic flux density optical collectors and develops a new technique to quantify correction functions for angular imperfections. The detailed responses of a pair of 2 pi optics used for a Zeppelin and a second pair used on the HALO aircraft. They are examined individually as downwelling ground based detectors and in 4 pi pairs as they perform on the aircraft. These optics show significant differences in the azimuthal and zenith responses and thus differ in the corresponding correction functions.

I commend the authors for the rigorous and exhaustive examination of the optical collectors and the deep analysis to correct imperfections on their specific measurements. The tools developed provide a resource to the community to improve actinic flux measurements. Importantly, these corrections do not rely on knowledge of the atmospheric environment (e.g. clouds, aerosols, etc) but rather on the relative changes in the measurements themselves.

This work provides a call to all groups with measurements of actinic flux to consider optical accuracy and the impacts of the non-ideal response. While this may not be practical for all groups, I suggest experts in the community may support more detailed optical analyses.

That said, the corrections are not always large and depend on measurement requirements. The authors note that the magnitude of corrections depends on the optical quality and measurement geometries. Downwelling optics for ground-based UV measurements (e.g. spectrometers or filter radiometers) may have relatively minor corrections (e.g. Figure S28), particularly with lower surface albedo. Total actinic flux from optical pairs with a sufficiently accurate 4 pi response also have relatively small corrections (e.g. Figure 11 and the discussion Section 7). The authors note that mechanical adjustment can improve hemispherical measurements, though corrections may still be necessary.

In addition, the primary purpose for such measurements is the subsequent calculation of photolysis frequencies. The impacts of the corrections on photochemistry are reduced and the authors note this is beyond the scope of this work. Nevertheless, it is an important consideration in the larger picture. Most photolysis frequencies are driven by UV wavelengths and the UV corrections in this study were relatively minor. In addition, photolysis frequency calculations include large molecular parameter uncertainties that typically exceed the measurement uncertainties. Nevertheless, better measurements are in fact, better, and molecular uncertainties may be reduced in future studies. At longer wavelengths, the uncertainties were more significant affecting the photochemistry of NO3 and other molecules.

The paper is well written, relevant and fits well within the scope of AMT.

Minor comments and revisions:

Line 1 and/or line 28: UV/VIS to “ultraviolet and visible (UV/VIS)”

Lines 114-5: I do not see a description of FLT, FLV and FLN. Nor do I understand the acronyms. Figure S7 does give some information but does not distinguish between FLT and FLN.

Line 228: Typo. Change ‘was’ to ‘were’

Figures 7, 8: The paper jumps between Zeppelin and HALO configurations. Be sure to distinguish the platform in each figure caption for clarity. In these two plots they are the Zeppelin optics.

Line 358-60: Should this refer to Figures S4 and S5. Figure S11 does not show the azimuthal variability. Also, the variability of the nadir sensitivity looks to be about 5% based on the error bars in Figure S5. That does not seem to qualify as “small”. Perhaps note it is an exception here.

Figure S5: Just a comment: The bottom optic for the Zeppelin seems to be a significantly inferior optic. Perhaps that is why it was placed in the less consequential upwelling position. Could this be improved through mechanical adjustments? If so, the corrections would be more in line with the other optics. In addition, both Zeppelin optics seem to encourage crosstalk as they show 80% sensitivities at 90 degrees. Optimizing the angular response is a balancing act but perhaps these could be adjusted to be closer to 50%.

Lines 370-380: I think this is an important place to note explicitly that the impact of the large upwelling correction near the surface has only a small impact on the total. Thus, the impact on total photolysis frequencies would be small.

Figures 11 and S27: The grey bars for the Cs cloud in these figures are deceiving. The cloud ranges from 10-12 km but is shown as a thin line at 11 km. I also suggest adding a point at 12 km to show the correction from the bottom to the the top of the cloud.

Line 374: “Because of insufficient,…”. This line is vague. If I am interpreting the meaning correctly, you could modify the end to, “Because of insufficient field-of-view limitations of the bottom receiver, significant cross talk to the upper hemisphere occurs and the Z(NH) are generally greater than unity”.

Line 410: Typo. I think you mean “contain the uncertainties”

Line 420: Typo. “final”

Line 504-5: The main source of the increase in upwelling flux from 300 to 600 nm is the orders of magnitude increase in the solar spectral shape. The albedo effect is secondary.

Figure 17: The error bars show the uncertainties from the corrections and calibrations. I was disappointed not to see the impact of the corrections alone as that is the point of the entire paper. It should be shown in this figure (or another) and well explained in the text. The fact that they may be relatively small (especially on this HALO flight) is important.

Section S3.4: Note that the relative contributions apply at 400 nm. That is stated in the figure captions but not the text.

The authors determined the angular response of optical inlets for measuring actinic flux densities and provided corresponding correction functions for ground-based and airborne applications and a wide range of atmospheric conditions. Their approach is based on detailed laboratory measurements combined with extensive radiative transfer calculations. Overall, the manuscript is well suited for publication in AMT. However, some comments might be considered before.

General Comments

The authors give a lot of details to demonstrate what they did. The supplement material even extends their efforts to show everything that was made. This brings me to the main general comment. The reader might be overloaded by the number of plots and information specific to author’s optics and applications. Therefore, the authors should consider limiting their plots to those that show significant differences. For example, it is not necessary to show the results for the four selected wavelengths for all investigations.  

Specific Comments

P2L46: “the accuracy of measurements in the UV-B range” – I think it is rather a question of sensitivity what is meant here. Technically, the accuracy includes also the uncertainty due to a non-perfect angular response.

P2L52: Maybe also refer to Jäkel et al. (2007) (https://doi.org/10.1175/JTECH1979.1) who discussed the stray light correction for a similar instrument.

P3L64: “…, owing to the greater importance of upward radiation, …” Why is it of greater importance?

P4L96: “of typically high spectral radiances in both hemispheres.” This is not necessarily valid, e.g., flights performed over land or open water under cloud-less conditions show a low contribution of upward radiances.

P7 Fig3: I’m wondering if this could be combined with the cross-section plots shown in Fig. 4.

P8L172: “Because of the rotational symmetry of the receivers, dependencies on azimuth angles are minor.” From the contour plot (Fig.3b), I would estimate a distinct variation of the angular sensitivity in azimuth direction (for 60° polar angle). Is this considered as minor dependence?

P8L176: “The dependencies on polar angle and the wavelength Zp dependence are slightly different for the different receivers.” Please give the range.

P11 Fig5: Here, azimuthal averages are plotted. Think about to show also the corresponding standard deviation as in Fig. 4.

P12L200: “In panel (b) of Fig. 6 relative sensitivities were multiplied with sin(ϑ) to account for the solid angle contributions consistent with the ϑ-dependent areas in the projections of Figs. 3 and 5.” I’m not sure if this step is obvious for the reader.

P13L242: “The applied ground albedos are based on literature data.” Please give reference. Same for the cloud settings (L248).

P14L243: ”… considered a normal ground albedo” Maybe it is better to call it a “default albedo” for your study. Same for the aerosol optical depth (L249).

P14L263-P16L285: The authors present the simulations of the diffuse radiance field for cloud-less conditions. I would rather prefer to see a direct comparison to the more interesting cases that are shown in the supplement.

P15 Fig.7: Could be combined with Fig.8.

P16 Fig.8: Is it reasonable to give azimuthal means here, since the distribution for the downward component has such a large azimuthal dependence? Think about a plot showing the principal plane direction instead.

P20L349: “clear-sky corrections”: Does clear-sky corresponds with cloud-less conditions?

P22 Fig.11 / P23L390: “solar azimuthal position”: Maybe use the term relative azimuth angle instead.

P25 Fig.12: The number of dots should be 3 x 5 x 3 for this scenario, but it looks like less.

P27L474: “altitude-interpolated coefficients”: Do the atmospheric profiles of e.g. temperature and pressure have an effect on the altitude dependence?

P27L490: “A detailed description of the correction procedure is given in Sect. S7 of the Supplement.” Maybe it is useful to provide a schematic that illustrates the correction procedure directly in the main manuscript.

P32L552ff: The paragraph partly contains advices which are obvious (e.g., “if measurements are made on a pavement or artificial platform in an area dominated by vegetated ground, measured upward flux densities can be misleading.”

In my opinion, it goes beyond the scope of this manuscript to go into the question of how to measure actinic radiation. Here the authors could shorten the text.

P35 Conclusions: I would suggest to give a final quantification of the corrections to illustrate their necessity.

Technical Comments

P2L32: lifetimes - lifetime

P8L173: “are obviously invisible” - not visible ?

P9 Fig.4, P11 Fig.6: Please revise the legend. The dashed line is not shown as dashed line there.

P9l192: “in panel (a)” - give also figure number

There are several figures without labeling the four panels.

P28 Fig.14: Maybe adjust the range of the x-axis (zoom in, e.g. 06:00 – 10:15 UTC). The date is not required as x-axis label.  

P29 Fig. 15: Use the same y-scale if appropriate. It helps to make the plots more comparable.