Overall, the manuscript is of good quality considering writing and technical contents. I belive that the presented methodology, approaches and results are of interest for the scientific community, aiming for and significantly contributing to a standardization of radon flux measurements.

Despite this, I believe that minor revisions of the manuscript and adressing of minor technical comments below can improve this work further. While there are some detailed technical comments/questions on the methodology below, these are presumably not critical and can be easily addressed by the authors.

Comments on the contents:

• L74: It is stated that calibration of Radon flux monitors requires a calibration exhalation bed type facility. I understand that this is the ideal situation. It seems from later given equations/methodology though, that the derivation of the flux boils down to geometric factors and the calibration factor of the Rn-monitors used inside of some sort of accumulation system. Can it not therefore be done simply by calibration of the Rn-monitors with conventional reference atmospheres and calibration of the geometrical components of the flux system – and if not, why? Have you checked if such a route would lead to similar results compared to the exhalation bed facility? (I presume that the geometry of the flux system can be easily determined to relatively good accuracy). You state also, that for the INTE_Flux, the employed Rn-monitor was already calibrated previously in L358. Also, this seems to be the way that the exhalation bed values have been characterized in the first place. I think it would therefore be beneficial to give some more explicit info on the significance of the exhalation bed (is it more for proficiency testing? What is the advantage of the EB? Also see comments about the Conclusion section)

• L123-L124: You stated before that the emanation factor is given as the ratio between Rn activity that escapes from grains into the pore-space to the total Ra activity. Here, you state that it is given by the fraction of Rn activity escaping (i.e. that leaves from the pore space to ambient volume) to the Ra activity (Eq. 2). It appears to me that this is only valid for negligible bulk volume/masses, i.e. “small” enough or “well enough” distributed samples (as you state in L128). I suggest adding this information to the introductory text of Eq 2, i.e. in L122.

• L155/Eq 9: I believe the lambda without any subscript (in the denominator before the brackets) should be lambda_eff. (The formula in the brackets formally comes from a convolution integral over the non-time dependent flux term in the solution of associated ODE of the volumetric Rn activity in the “chamber of known volume”. This convolution integral should contain the kernel exp(-lambda_eff * (t – tau)) and be wrt. dtau, so it appears to me that this should be lambda_eff in the denominator).

• In L187 and L188 you state that there is no “statistical difference” (between measured and theoretically/semi-empirically derived exhalation rates) between theoretical and experimental approaches given in other work, which I agree is indicated by their data. However, please include a statement at what uncertainties this is. It appears to me, that the theoretical approach in this cited work has uncertainties around 20 %. The way this is written here esp. considering L186 (lack of value/uncertainty for theoretical approach) suggests that there is some sub 2 % agreement between theoretical models and experimental values, which can not be concluded from this other work.

• L185 why are these experimentally determined values considered to be an “estimate”, despite the very low cited uncertainties? Also appears at several different points in the text, i.e. L213, L438, 448. Isn’t the experimental method a “determination” rather than an “estimate”? (I agree on the theoretical one being an estimate though). Since the “experimental method” was what apparently determined the “reference value”, I don’t think this should be called an estimate, especially not in L448. Please also give indication more explicitly there, that this is the “go to” value for the rest of the work (i.e. the one that was used for the actual calibration in L529).

• L241 following, Eq 10: It appears to me that this “linearization” approach is not a main contribution of this paper, but no references are given (also used/derived in cited work by Gutierrez-Alvarez et. al.). I understand that in practice, lambda_eff is not known and therefore, Eq. 9 is linearized for it to “cancel out” in Eq 10. Please include this information explicitly, because otherwise, it is not clear why the equation needs to be linearized in the first place. (i.e. if lambda_eff was known, Eq 9 would already be linear wrt. to 1-exp(…)). Is there an estimate on the error limits introduced by the truncated terms (I presume this scales with lambda_eff*t), maybe also elsewhere, which can be cited? (Error estimate can also be in L304 where specific integration time is mentioned)

• L266: You state that the Radon monitors influences the “system response” time by diffusion. This is very true, but another factor is for the (not rare) systems which measure the Rn progeny (as a proxy) rather than the volumetric Rn. For those, the response characteristics are even way different because of the progeny ingrowth, but since their detection limits are typically way smaller, it may be tempting for future work to use such a monitor. I think it would be beneficial information to include here explicitly that the approach can only reasonably be applied to Rn monitors which measure actual volumetric Rn.

• L278: Shouldn’t the required detection limit depend on the volume of the accumulation chamber and be therefore expressed in terms of a proportionality factor?

• L382, 383: Celaya et. al. is not in the reference list and therefore, I could not look it up there: Please include which gamma-emissions have been used for determining the Ra-226 activity). Since it is stated that the sample was equilibrated beforehand, I presume that the progeny emissions (e.g. 352 keV) have been used rather than the 186 keV Ra-226 line. I assume that is due to interference with Uranium, which could have been corrected for using its other emissions. Please include info on this explicitly, otherwise it is unclear why the sample had to be sealed beforehand.

• L391: So, was the stated uncertainty of epsilon just determined as the standard deviation of the three identical experiments, or have other uncertainty contributions been included/considered too? Which of these two is the major contributor, i.e. is the “across samples” variation comparable to the uncertainties theoretically derived for each of the samples? If it isn’t, what might be the cause for this (also aiming into the direction of why the three identical replicas were needed in the first place)? I consider this important also because the determined epsilon together with the relatively high uncertainty is a big reason to why you can state in several other places (namely e.g. L505, 506) that all values are consistent with each other wrt. to given uncertainty, especially considering the semi-empirical/theoretical approach (which seems to me only compatible because of this high uncertainty).

• Eq 11, L395, L396: This refers to figure S5 of the supplement rather than figure S6. Was the correlation of the lambda_eff and phi (both appear in the associated equation) resulting from the regression considered in the uncertainty propagation for epsilon (or is this rather just stated as the standard deviation, see previous comment)?

• L429 please explicitly state the “water saturation values” for which this applies

• Table 2: It appears to me that the “experimental approach” to determine the exhalation rate was carried out using Eq 10 (L448). However, the budget given here contains some symbols that are not included in Eq 10. What is their significance?

• Eq 15, 16, 17: Am I misunderstanding or are these equations not considering the ingrowth-/decay of Rn in each of the compartments. If so, why not? Is it because this is insignificant on the considered time-scales and since these equations are just to “understand which data should be analyzed” and to thus keep the modeling simpler? If so, I think this should definitely be mentioned in the text.

• How is “F_Th_AF” in L504 determined? Is this just an estimate “by eye”?

• Conclusions section: I believe this should be extended wrt. to several points, especially considering the exhalation bed that takes up a significant portion of the work. It is not mentioned/clear if this exhalation bed is now good/better than others. Further, it is not clear what the purpose of the “theoretical approach” is in the first place, since the results are not discussed, put into perspective nor used in any further calculations in the calibration, correct? Please discuss this aspect of the work, especially, in my opinion, the stated agreement between the theoretical and the actual values considering the relatively high associated uncertainties. Also, the conclusions seem to be somewhat more connected with the overarching goals of the project rather than the main parts (calibration, transfer standard, exhalation bed) of the presented work. I suggest putting more focus on the actual main contents of this work,e.g., why is the ANSTOFlux a reasonable transfer standard compared to others etc.?

• L635 and Figure 8: I think it is surprising that the model driven by the ERA5 data produces considerably “better” results than the model driven by the actually determined (i.e. local) parameters. Wouldn’t one expect, for a correct/working/accurate model, to work best on the best (i.e. the local?) data?

Beyond scope comments:

• Would it not be possible to derive an analytical equation for the regression analysis from Eq 15, 16, 17 in order to include the now excluded data? (i.e. similar to the “blue values” in Figure 5, but with the reference values as F and fitted other parameters?)

Formal comments:

• Line 28 (L28): I think, given the definitions used in the paper, that “exhalation” would be want is meant here instead of “emanation”
• L42: I suggest keeping the germs “noble” and “gas” together (i.e. “radioactive, noble gas”)
• L85: I think “equipped with [..]” instead of “provided with [..]” is better wording
• L117: A reference for the Rn decay-constant should be given, in my opinion.
• L239: “ionization chamber” rather than “ion chamber”
• L326: Please include “nominal volume V_D” or include an uncertainty (since many digits are given).
• L341: I think information on the specifics of IT (i.e. raspberry Pi ethernet, “Bitwise” client) can be omitted, since it serves no further purpose.
• L437: I believe it should be S6 rather than S5 here.
• Figure 5: Please correct typo “withing” in the labels.
• L595: Typo in “properties”
• L620: I believe this should be “purposes” rather than “proposes”
• L622: “To” rather than “for”
• L626: “properties”

The manuscript “Characterizing the automatic radon flux Transfer Standard system Autoflux: laboratory calibration and field experiments” porposes a new standard method for radon flux calibration which is an important contribution for those who works in this fiel mainly due to the lack of standardization. Writing is of good quality and provide sufficente datais for those who wants to implement this metodoloy in their on laboratories.

I consider that just minor revisions of the manuscript are necessary for the publication

L396 - A typical measurement result is shown in Figure S10 – in fact is Figure S8

L438 - Figure S8 of the supplementary - in fact is Figure S9

L484 - Figure S13 shows the three main volumes - in fact is Figure S11

Athought discussed in the item 3.4 my main concern is due to the fact that all experiments were done with high Rn activity concentrations, whch will not be found in enviromental studies and I agree that more studies must be made in order to verify the valdity of the proposed model.