The authors compare gravity disturbance results from a strapdown, iMAR IMU and a gimbaled cold atom absolute gravity meter, both collocated in the same aircraft, flown over areas with previous gravity data.  By looking at crossover (and repeat line) statistics, they derive optimal filter lengths (long enough to smooth out sensor noise, but short enough to capture spatial changes in the gravity field).   By spectrally combining the two instruments they achieve the best statistics, indicating the absolute meter is really providing a long term "basis", while the iMAR robustly handles short term signal changes.

My first reaction is that this seems to be much more than just a typical "data description"!  There is a very detailed description of the analysis techniques:  flowcharts, Fourier transforms in combining the two sensors, etc.  My suggestion would be to embrace that and offer some more aggressive conclusions:  crossover statistics are a useful tool in designing optimal filter lengths (whereas repeat lines might not be, and that combining the two sensors really does take advantage of each instrument's strength to provide a better data set.

Also, the paper starts out indicating that repeated surveys would "enable investigation of mass variation, assuming the accuracy is sufficient."   The paper never answers this.  Is 0.5 mGal enough for Vatnaokull (probably?), or 1.2 mGal for Nuuk (maybe?).

All that said, and after addressing a few detailed comments (below), I recommend this paper be published in Earth Systems Science Data.  The techniques and results will be of wide interested to the airborne (and other mobile applications) gravimetry community.  My congratulations to the authors on a very nice experiment!

Detailed comments:

--------------------------

Line 11.   "..from *an* airborne gravity campaign..."  

L14.  Suggest adding "...only measure the *relative* variation of gravity...".   And technically, by "gravity" you should point out that you at least mean "acceleration of gravity" or, preferably, "specific force".   This comes up later, but it's probably best to be precise from the beginning.

L62.  Without knowing which way north is, it is not so easy to identify the "east-west flight line".  Recommending adding a "North symbol" to the figure.

L70.  In addition, the GIRAFE should have other advantages, true?:  No need for airport ties, each measurement is independent of previous measurements (neglecting any co-sensors), etc.

L90.  For completeness, is it possible to report on the distance from the antenna (and thus gravimeters) to the aircraft's center of mass?

L133.  Perhaps explicitly state that delta_g is the gravity disturbance here (rather than "introduce" it in the next section..).

L158.  Now I am intrigued:  can you say more about not using the IMU/GNSS solution to determine h_doubledot (but it *is* okay to use it for h)?

L162.  This is the first we have seen (x,y,z) components.   At this point (probably earlier where different inertial frames are first mentioned...) we need a figure at least linking xyz to alpha and beta.

L176.  This sentence is a bit confusing.  Suggest something like "magnitude of the expected gravity signal variation at altitude..."  along with "... a heavy low pass filter..." emphasizing that the SNR is tiny.

Table 2.  I suggest pointing out in the caption that delta_g is the disturbance.  It could be confused for the uncertainty on g.  Note that none of these values has an associated uncertainty (and at least the gravity and disturbance really should).  Also, how were the apron values determined?  Using GIRAFE or a "classic" absolute meter?   If the latter, was the GIRAFE in agreement at the apron site?

Figure 6.  Recommend swapping the two images left for right:  It is more customary to describe the left plot before the right.

L215.  The (unnumbered...) equation appears to be nonsensical:  a = b - a.  The residuals have the same symbol as the reference field.

L277.  As above, recommend putting a north indicator on the map.

Figure 10.  Suggest to rewrite the caption as "Mean, standard deviation, and RMSE of the cross-over differences as a function of filter width...".   It's not clear at all from the figure nor until very late into the caption sentence what's actually being plotted!

L.280.  Note that the "superior coverage" of the iMar is technically true, but actually quite subtle and not obvious at first glance.  Please quantify the difference.   The same is true later for the Nuuk Fjords on L317. 

L291.  Re: the mean plot in Figure 10...  The fact that iMar mean crossover values are almost all below 0, with the worst mean crossover actually happening at the best/chosen FWHM value (which makes sense given the minima in both stdev and RMSE...) needs more discussion.  I assume it just means that one (or more) of the cross lines has a bias.  Can the iMar suffer a "tare" relative the tie value at the airport?  What about attempting to match a long wavelength, mean value to a global model to remove the bias?   Anyway, more discussion is suggested.  (Especially as this seems to be the final conclusion.  See L351)

Section 4.2  This little section does not seem to add too much to the discussion.  I understand that the ground signal is expected to have changed, but is it possible to at least quantify the differences between measurement and prediction?  One is left to "guess" at value by looking back and forth at 11a and 11b.  The size of the change would be interesting to compare later with the accuracy/resolution of the instruments (for example).

L327.  Again, a north indicator on the maps would help explain the discussion.  Apparently the east-west repeat line is the "lonely" line in the middle (?).  

L332.  There are a few small English mistakes here and there, but "argumented" should definitely be "argued".  

L33.  A choice of 40s seems reasonable when looking at the crossover results, but the east-west repeats would indicate something more like 90 or 100 (especially for the GIRAFE).  This probably needs more discussion.  Is the conclusion that east-west repeats are not a good method for estimating optimal filter length?   This would be important, because in rugged terrain, one might expect that different filter lengths could "smear" a feature in one direction, and minimizing that would help in filter length choice.

Conclusion.  As mentioned before, if the intent was to only describe public ally available data sets, this paper could have been much shorter.   I be live though that the detailed description of the analysis will be of wide interest to the dynamic gravity community.  If some of the ideas here - using crossover statistics to design optimal filters (and perhaps the poor utility of repeat lines for this purpose) - are already described elsewhere, they should be referenced.   But if they are novel , go ahead and list them as useful conclusions.  And again, knowing if the instruments are "good enough" for the geophysics happening in Iceland and Greenland would also be very interesting.

This paper presents the data obtained during the airborne gravimetry campaigns conducted in Iceland and Greenland in 2023. Notably, the campaigns employed not only a conventional relative gravimeter (the iMAR strapdown gravimeter) but also a cold-atom absolute gravimeter (the GIRAFE quantum gravimeter). The acquired data were evaluated through cross-over analysis, which confirmed that the individual datasets and their combined solutions achieve an accuracy of 1 to 2 mGal. Furthermore, the cross-over analysis highlighted the respective strengths of the two instruments: the GIRAFE absolute gravimeter demonstrates resilience against systematic bias, while the iMAR relative gravimeter excels in spatial coverage due to its robustness in handling disturbances.

The dataset presented in this paper is both unique and valuable, as airborne gravimetry employing quantum absolute gravimeters remains rare globally. The authors' consideration to structure the data into levels ranging from Level 0 to Level 2, catering to diverse user needs, is highly commendable.

The manuscript is well-written and appears to provide the necessary information required for a data paper. I believe it aligns well with the scope of the ESSD journal and is a strong candidate for publication.

Below, I offer some minor comments for consideration.

---

L19, L21, etc.: According to the ESSD submission guidelines (https://www.earth-system-science-data.net/submission.html), the format for in-text citations should follow these rules:

- For example, "(Jensen(2024))" should be formatted as "(Jensen, 2024)."

- Similarly, "(Bidel et al. (2020), Bidel et al. (2023))" should be revised to "(Bidel et al., 2020; Bidel et al., 2023)."

L80: It would be beneficial to include information about the model of the GNSS antenna as well.

L89-90: How was the lever-arm estimated? It would be helpful for readers to include a brief explanation of the estimation method, along with a reference if available. Additionally, since the accuracy of the lever-arm affects the absolute accuracy of the observed gravity values, it would be helpful to include their estimation errors, if available.

L150: It would be helpful to include the version of the ITRF as well.

Table 2 caption: "iMAr" should be corrected to "iMAR."

Figure 6: The title of the scale bar in the right panel should be changed from "x-over" to "cross-over" for consistency.

Figure 6 caption: It would be more intuitive to describe the figure from left to right.

L215, 219, 223, 226: Equation numbers should be assigned to these formulas as well. Additionally, the same symbol is used for "GGM-reduced gravity anomaly" and "GGM-derived gravity anomaly". This should be corrected.

L220, 248: It would be helpful to specify which covariance function model was used in these LSC calculations. 

L235-236: The sentence "The terrain effect computations take into account the density of ice (0.92 g/cm³) is different from the conventional rock density of 2.67 g/cm³" is somewhat difficult to read and could be revised for better readability.

L239: It would be helpful to include an explanation of how the error in upward continued gravity disturbances (2-3 mGal) was estimated.

L280, 317: While it is true that the iMAR measurements cover longer survey lines, describing this as "evident" might be slightly overstated. I suggest using a more moderate expression here.

Figure 10 & 14: The cross-over difference statistics for iMAR appear to change smoothly with filter length, whereas those for GIRAFE seem to lack this smoothness. What specific characteristics of GIRAFE might contribute to this behavior? If there are any potential causes, it would be beneficial to include them in the text.

L342: It would be helpful to provide an explanation of the method used to hybridize the GIRAFE and iMAR data.

Figure 16 & 17: I recommend creating a table to summarize the cross-over difference statistics for iMAR, GIRAFE, and the combined solution at the optimal filter lengths. Such a table would provide a clear and concise comparison for readers.

---

Review about the paper „ Airborne Gravimetry with Quantum Technology: Observations from Iceland and Greenland” submitted to Earth System Data Series by Tim E. Jensen et al.

In this publication, the authors report on two flight gravimetry campaigns over Greenland and Iceland in 2023, where they simultaneously deployed a novel quantum gravimeter on a gyro-controlled platform together with a modern strapdown gravimeter. The installation of the measuring devices in the aircraft, the execution of the flights and the processing of the data are described in detail. The authors compare the measurement data obtained with the two different sensors and combine them in an interesting, intentive way. In addition, the data products are evaluated using cross-over analyses and compared with other terrestrial gravity measurements and data from previous flight campaigns. All in all, in my opinion, this is a convincing work, which is of particular value in that the authors have also published the herein obtained data freely available to the user community. This should give interested readers the opportunity to familiarize themselves with the two different modern sensors using the raw and processed GNSS and gravity measurement data and to understand and verify the authors' results. It should be noted that raw gravimetric measurement data of this kind are rarely published for the general public to this day. In this respect, the authors follow the international FAIR principles for the description, storage and publication of scientific data, which I particularly appreciate.

I have read the paper carefully. I recommend to publish this paper after some minor revisions according to my following comments:

• It seems for me (from figure 3), that your iMAR iNAT-RQH strapdown gravimeter was operated without a temperature-stabilising housing. What about the drift behaviour of the strapdown gravimeter when running it in the cabin of the aircraft in such a way? From Johann (2019) it is known that a thermal IMU calibration may be necessary here. Can you please comment on this item in the paper? Which drift values did you estimate for the iMAR iNAT-RQH instrument?
• Furthermore, in Johann (2019) an empirical heading-dependent correction is described. In the meantime, it is known that such course-dependent distortions can occur in iMAR strapdown gravimeters data due to the influence of the Earth's magnetic field, depending on the type of the used accelerometers. Can you please comment on this topic in your paper?
• In sections 3.6 and 4.5 you describe a combination of the data of both sensors based on a FFT approach. I see a little contradiction between formulae (9) and formula (10): While the indices k and m in (9) start with 0, n in formula (10) starts with 1. I suppose this is a mistake, n should be between 0 and K, isn’t it?
• In section 4.5 you write: “Such hybridization is relevant also for the definition of future satellite missions.” I do not understand this statement and I fear it’s wrong. From my point of view, such kind of hybridization based on a FFT approach is hardly feasible in case of satellite gravimetry missions. Satellite based gravity field estimation is based on completely other data processing techniques (for instance least squares adjustment of huge numbers of data)
  
  
  
  

Further comments/recommendations:

• Page 1, line 11: Please correct: “This manuscript reports on the data available from and airborne gravity campaign carried out in summer 2023.” into  “This manuscript reports on the data available from an airborne gravity campaign carried out in summer 2023”
• Page 1, line 14: I suggest over-emphasising this statement somewhat: “Unlike classical technologies that can only measure the variation of gravity from an aircraft…” → “Unlike classical technologies that can only measure the relative variation of gravity from an aircraft…”
• Page 3, figure 1: I recommend to indicate the Snæfellsjökull volcano in the map as well
• Page 11, line 181 ff: The sentence “… the iMAR strapdown sensor are corrupted by the presence of systematic errors in the gravity sensor” sounds a bit like that the strapdown sensor is faulty. I recommend such a formulation: “Following the processing described in the previous section, the relative gravity disturbance estimates derived from the iMAR strapdown sensor need to be corrected for the therein contained systematic errors.”
• Page 11, table 2: To avoid misunderstandings, please change “Height” into “Ellipsoidic height” in the header.
• Page 12, figure 6: Could you specify over which location the plotted tracks were taken?