General comments:

The manuscript presents the Broadband AirCrAft RaDiometer Instrumentation (BACARDI), which is designed for the observation of incoming and outgoing broadband shortwave and longwave radiative fluxes from the surface into the lower stratosphere using thermopile-based radiometers on board of the High Altitude and Long Range (HALO) research aircraft. The manuscript describes corrections for the temperature dependency of the sensor thermopile sensitivity, the sensor response time and the attitude of the aircraft. In addition, a new method is introduced to correct for thermal offsets. The correction uses the measured body temperature of the sensor to calculate the thermal offsets. The corrections were evaluated using measurements from the Elucidating the role of clouds-circulation coupling in climate (EUREC4A) field campaign. Vertical profiles of observed radiative fluxes up to 10 km altitude demonstrate that the proposed thermal correction reduces the thermal offset from original 10 and 20 Wm^-2 for pyrgeometers and pyranometers, respectively, to below 10 Wm^-2. In addition, comparisons with radiative transfer calculations for the shortwave wavelength range confirm that the corrections reduce the uncertainties substantially.

The radiation balance in the atmosphere plays a fundamental role in the climate system. However, in-situ observations for the validation of model outputs and satellite products are limited to a few profile and in-situ measurements taken from aircraft and balloons. In addition, these measurements suffer from higher uncertainties due to the challenging atmospheric environment with rapidly changing temperature conditions and the attitude of the platform which have a considerable impact on the accuracy of radiation measurements. The radiometer set-up and the correction methods presented in this manuscript will definitely contribute to a better accuracy and reliability of such in-situ profile observations.

The manuscript is very well structured and very clearly written. The methods are thoroughly described and the results reasonable. The literature has been selected and cited carefully. Graphics and tables are clear and the captions self-explanatory. This work is a very interesting and a highly valuable contribution to the atmospheric science community and is in my opinion absolutely suited for publication in AMT. I recommend publishing with minor revisions and technical corrections.

Specific comments:

• Corrections methods and their evaluation are thoroughly described for the shortwave and longwave. In addition, a comparison of the observations with radiative transfer calculations for the downwelling and upwelling shortwave radiative fluxes for horizontal, circular flight patterns has been presented. Such a comparison would be also highly valuable for the (downwelling) longwave in order to estimate the reliability of the longwave observations and to study potential effects which may become relevant in the longwave on such flights. For instance, the fraction of the direct solar beam above the cut-on of a pyrgeometer, which is at about 4.5 μm for a CGR4. This unintentionally observed portion of the direct solar beam depends on the water vapor content (and thus altitude) and the solar zenith angle and hence may exceed 5 Wm^-2 significantly on such flights (e.g., Marty, 2000). In addition, the dependency of the pyrgeometer sensitivity on the water vapor content, which is estimated to be about 5 Wm^-2 in cloud-free conditions, may also be considered in such applications (e.g., Nyeki et al., 2017).
• Would it possible to calculate an uncertainty budget for the BACARDI package or at least to give a conclusive uncertainty estimate for the individual components of the observed radiative fluxes and the net radiation?
• It is indicated that the thermal offset correction coefficient β of the upper and lower pyrgeometer is more consistent compared to the upper and lower pyranometer due to the position of the pyrgeometers in front of the pyranometers with respect to the flight direction, which allows the pyrgeometers to be ventilated more effectively. Would it possible to place the pyranometer to the side of the pyrgeometers to further reduce thermal offsets or impedes the mounting system of BACARDI or limited space in the fuselage such a setup?

Minor comments and/or technical corrections:

• Line 32: may use “… by radiometers, … pyranometers … pyrgeometers”
• Line 55: may use “Actively stabilized pyranometers, …”
• Line 83: may use “The radiative energy budget of a broadband radiometer”
• Lines 129/140: In my opinion, the sensitivity is normally given in units of V W^-1m² (see line 189). Hence, the stated unit Wm^-2 V^-1 refers to the reciprocal of the sensitivity à may use “…adjusted reciprocal of the pyrgeometer/pyranometer sensitivity…”
• Line 157: delete one “the”
• Line 313: “… depends…”
• Fig. 5: Is there an indication for the cause of the “outliers” in Fig. 5 (grey points)? Are these the same datapoints for the shortwave and longwave? May give a short statement in the text.
• Line 346/347: I would replace “… a few W m^-2” by “…to values below 10 W m^-2” (as in the abstract). For the downwelling shortwave flux, values are rather between 5 and 10 W m^-2 above 6 km but also partly near the surface. Only in the upwelling shortwave and in the downwelling shortwave up to 6 km values are a few W m^-2.
• Line 416: Delete either “the” or “a”
• Line 432: I do not understand the expression “… false detection of clouds above the aircraft…”. May rephrase, e.g., “… caused by cloud contamination above the aircraft in the filtered dataset, …” or similar. Significant lower observed irradiances with respect to the calculated fluxes are either due to real clouds above the aircraft or a not properly corrected misalignment of the sensor.
• Line 433: “… conditions of high solar zenith angles, …”
• Lines 413-456: I got a bit confused here: Fig. 9a is presented and described in lines 413-423. Then Fig. 10a and 10b are described in lines 424-448. Finally, you go back again to Fig. 9b in lines 449-456. It might be easier for the reader, if the description of Fig. 9b (lines 449-456) was placed right after the description of Fig. 9a in line 423. However, you may have good reasons not to do it.

References:

Marty, Ch.: Surface radiation, cloud forcing and greenhouse effect in the Alps. PhD-Thesis, No. 13609, Swiss Federal Institute o Technology (ETH), Zurich, 2000.

Nyeki, S., Wacker, S., Gröbner, J., Finsterle, W., and Wild, M.: Revising shortwave and longwave radiation archives in view of possible revisions of the WSG and WISG reference scales: methods and implications, Atmos. Meas. Tech., 10, 3057–3071, 2017. https://doi.org/10.5194/amt-10-3057-2017.

This paper describes the physical set up and processing methodology for a broadband radiometer measurement system on the HALO research aircraft. Both the engineering design and the processing are carefully done, tested, and documented. In particular, a derivation of the theory behind thermal offset corrections in quickly varying temperature conditions such as ascent and descent is given and applied to the data. Downwelling shortwave data is also corrected for attitude in flight using methods documented in the literature. The manuscript is well written, the methods are scientifically sound, and the work describes a high-quality radiation measurement system for a specific aircraft. This paper documents measurements that will be of great value to the scientific community for measuring radiative heating rates. I recommend this manuscript for publication after the authors respond to the minor points below:

1. The abstract states that the correction function “depends on the mounting position of the radiometer on HALO.” Please clarify what is meant by mounting position, and how conclusive this statement should be as several statements were made in the paper related to the impacts of mounting position, some of which seemed fairly conclusive and others more as hypotheses. In section 3.3 (Figure 3), the larger differences in sensor temperatures between the pyranometers than pyrgeometers rather than the upwelling or downwelling instruments is explained as being a matter of “the internal sensor housing” which I took to mean inherent to the differences in the instrument construction rather than how they were mounted on the plane. In section 5, the coefficients of the upper and lower pyranometer are described to differ by a factor of 2, and this is attributed to differing airflow between the two systems given the slight tilt of the plane. Then later in section 5, the up and downlooking pyrgeometers had much similar beta values which was hypothesized to be because “the CGR4 sensors are placed in front of the CMP22s”. While there is a difference between sensor temperature agreement and agreement in coefficients for the corrections for dynamic offsets, they are related. I am curious whether the explanations that the authors give for these factors that cause differences in CMP22’s and CGR4’s thermal responses (position relative to airflow in section 5, and internal sensor differences in section 3.3) are perhaps related and how important the relative ventilation of the radiometers is thought to be in comparison to sensor differences.
2. I had a couple of questions about the practicality of the attitude correction method used. It seems to be sufficiently accurate, and the HALO aircraft remarkably level in most flights, so these are minor concerns that the authors shouldn’t need to address for the publication of this paper. But I still found myself curious about a few practical details. The authors state that only the direct beam should be corrected for, so this correction should only be applied to clear sky conditions. I agree with that, however, I didn’t understand how the data was determined to be clear or cloudy. In the test case, the profiles could be determined to be clear fairly easily by visual inspection, though I would imagine this would be a harder job for a full field campaign. Was this correction run at all times and it left to users to determine whether to use the corrected or uncorrected data or is some kind of determination made for a best estimate value? Also, as the correction was based on radiative transfer calculations using atmospheric profiles from drop sondes or radiosonde launches—I was curious whether these will always be available for all campaigns where the HALO flies?
3. My primary concern with the methodology is that it wasn’t clear to me in the text which thermal offset corrections are applied to the data (that derived in section 2.3 only, or also a correction derived in section 2.2). In Figure 6, after the dynamic linear-fit correction has been applied, there are still negative biases in the downwelling solar irradiance of 5-10 W/m^2 at night. I don’t see an adequate explanation for what this bias is. The reason given in lines 348-349, “caused by different uncertainties such as the radiometric calibration of the pyranometer”, seem quite hand-wavey and not satisfying to me compared to the careful work done elsewhere in deriving the corrections. A calibration error is multiplicative so shouldn’t give a bias at night. It seems more likely to me from the shape of that bias (larger with higher altitudes) that it is in fact related to a thermal offset (like that derived in section 2.2) that isn’t corrected for using the “beta” linear fit. The author’s state in lines 325-326 that a more complex multi-variate fit including Tref doesn’t improve the correction, and conclude that therefore the dynamic dome effect can’t be discriminated from the thermal offset. But they don’t show those results, and I still can’t help but think that the post-correction results in Figure 6 look like they are still impacted by an equilibrium thermal offset. Also, Figure 4 shows downwelling SW offset corrections even in level flights when the temperature doesn’t appear to be changing significantly, which implies that the static offset is taken into account in some way. So it was unclear to me whether the dynamic offset (beta) correction was applied to this data or a static offset as derived in section 2.2.

Minor comments:

1. Line 6: it would read better as “an efficient new method”.
2. Line 30: should it be “which can be measured directly”
3. Line 55: should be “Actively stabilized pyranometers”
4. Line 80: Section 6 is not specifically referenced in the paragraph about the structure of the paper. Did you wish to add that?
5. Line 92: should be “In the case of the pyrgeometer”
6. Line 108: I think rho_p should be rho_d in the rho_s*rho_p << 1 assumption.
7. Line 157: two the’s at end of the line
8. Line 244: should be “To enable maintenance”
9. Line 265: What does “one magnitude lower” mean? Does this mean one order of magnitude lower?
10. Line 313: should be “depends”
11. Line 416: The wording at the beginning of this line is unclear.
12. Line 508: should be “The data are used by Luebke et al (2022)”