Dear referee RC1,

Thank you very much for your valuable and detailed comments on our manuscript. We would like to incorporate all or most of your suggestions into the revised version, subject to the final decision of the editor. Here are our responses to your specific comments:

1. Comment: “L26: Refer to wet and dry deposition here.”
   • Response: We will refer to wet and dry deposition on L26.

1. Comment: “L35-37: In this sentence, it is put that it is difficult to estimate regional or national inventories due to large uncertainties in the NH3 field measurements, but the first two paper does not mention this, Wang et al. mention it specifically.
   
   There are indeed uncertainties in measurements, but also very high variability in emissions due to variability in conditions. Newer studies have shown that the uncertainty of model estimation based on a large number of field measurements are on the same order of magnitude as between measurement methods.”
   • Response: We agree. The first two references shouldn't appear here. They should be in L45 instead and will be removed from L35-37.

1. Comment: “Chamber methods
   
   There must be a distinction between ‘normal’ chamber methods as dynamic flux chambers or wind tunnels that measure continuously over time and the open chamber DTM that measure at a discrete time.
   
   L40-41: Pacholski et al. and Roelcke et al. use the DTM chamber, so the critique is justifiable here for DTM, but in my opinion that does not hold for the other chamber methods. Also, the underestimation mentioned here only applied for DTM in Kamp et al. and not for wind tunnels. Wind tunnels have often been used with acid traps where no interpolation is needed.
   
   In general, for the manuscript there is too little nuance on the chamber methods and how they can be used, and it should be very clear that DTM is a special case of a chamber method thus general terms cannot be used for DTM as a chamber method.”
   • Response: Yes, wind tunnels should be distinguished from other discrete chamber methods. We will expand the paragraph to include more distinction between chamber methods.

1. Comment: “It should also be clear that enclosure methods, including DTM, are not suitable for estimation of absolute emissions, but can be valuable when comparing e.g. treatments.”
   • Response: According to previous publications cited in this manuscript, the DTM was made suitable for measuring absolute emissions by measuring several times during the day and by applying a calibration equation using wind speed data at 2 m height. One goal of this manuscript was to verify that this was correct through laboratory testing of the detector tubes. We have therefore decided not to include such a judgment early in the introduction.

1. Comment: “DTM is mentioned as an alternative for remote areas without power, but other methods also exist (ALPHA samplers, absorbing sponges) that all integrate over time. Other alternatives should also be discussed to justify the use of DTM.”
   • Response: To our knowledge, alpha samplers have not been verified for use in small or heterogeneous plots. Also, alpha samplers and absorbent sponges require experienced laboratory personnel to prevent contamination during sample preparation and to ensure accurate measurements. We will add sentences to make both points clear to the reader and mention these methods in the revision.

1. Comment: “DTM gets a snapshot in time, and this is a critical point for this method making it very labor intensive to obtain a good time resolution in the measurements. This needs more attention in the discussion.
   
   Furthermore, the chambers are very small. There should be some discussion on how the chambers are placed in the field, e.g. after field application of slurry by band spreading or mineral fertilizers where there is a difference depending on the amount of granulate or slurry inside the ring.
   
   An assessment of the calibration to obtain absolute emissions should be added.”
   • Response: A comprehensive guide to the use of the DTM and proper fertilization is already available from Pacholski (2016), so we didn't include it in our introduction. We also believe that discussing the effects of chamber size and fertilizer application in the introduction would be outside the scope of our work, since our experiments measured the errors inherent in the Dräger tubes and chamber system material, not the calibration, application method, or chamber size. However, we will include a paragraph on this in the discussion.

1. Comment: “L44-46: Very much has happened and ammonia emissions have been measured in many field plots over the last 22 years (Bouwman et al., 2022). The present study should put into context literature published in recent years.”
   • Response: We will explore this further and incorporate the latest knowledge on uncertainties.

1. Comment: “L57-59: A calibration was applied, but as literature shows, there is ‘inherent underestimation of NH₃’. Please rephrase to make it clear that the calibration itself is uncertain and a subject that needs to be addressed more in the future.”
   • Response: Should we mention that the calibration has a range of uncertainty, or that it is uncertain that the calibration is correct?

1. Comment: “L90: A reference for the 10-15% should be added including how this is determined. In my experience there is some uncertainty related to the manual read of and how long time you wait after exposure.”
   • Response: The uncertainty here comes from the reference in L93 (Drägerwerk AG, 2011), which also mentions how it is determined. We will quote the reference again in L91 and also mention how it was determined.

1. Comment: “MGA⁷ is mentioned in the introduction, but CRDS is used in laboratory setup. Please adjust introduction. L127-128 + 133-134: CRDS is the reference, which is an ok choice, but it should be explained why and why it was not the MGA.”
   • Response: We will mention CRDS in the introduction. The main reason we used the MGA instead of the CRDS in 3.3 was that we did not have a mobile system for the CRDS, but we did have one for the MGA. We will mention this reasoning in the Materials and Methods.

Comment: “Also, please consider not to use MGA7 and G2103 as abbreviation as the names point to the manufacturer.”

• Response: “Yes, we will refer to the devices as CRDS for G2103 and MIRS for MGA7.”

1. Comment: “L209: Is the precision range for MGA correct? (0.04–1.80 %)“
   • Response: Yes, it is wrong in the text. We will correct this in the revision.

1. Comment: “Section 2.3: Add the actual temperatures. Also add to section 3.2.”
   • Response: We will add the temperatures to section 3.2.

1. Comment: “For the outdoor measurements. The air exchange rate is the same for DTM and MGA, but as I understand it from the description, air is pulled through the chamber for 60 minutes whereas the DTM was only placed there for 1 min. This should be justified as the release over time can change. A constant emission over 60 min must be the same as for the 1 min with the DTM.“
   • Response: The wait time for the MGA was necessary to ensure that response time was not a factor in the MGA measurements. Also, for the MGA measurements, we only recorded the last few minutes of the reading, not the full 60 minutes. We tried to ensure that the release over time was the same by starting the DTM measurements immediately after the MGA measurement within the same chamber system. We will make this clearer in the text.

1. Comment: “Figure 2 and Figure 3: Please consider using a different plot type. The individual points can be added for each concentration with a horizontal line for average with error bar or boxplot.”
   • We will change the plot type for Figures 2 and 3.

1. Comment: “Section 3.1: The SD’s for DTM are very different between measurements. Please elaborate more on this.“
   • Response: We will describe the difference between the measurements more clearly in Section 3.1.

Comment: “I’m missing the plots of linear regression, please consider including them as minimum to a document with supporting material.”

• Response: We will include them in the supporting material, as they don't provide much more important information than the numbers and Figure 2 already do.

1. Comment: “Section 3.2: I’m missing the plots of the response time for the different materials, please consider including them as minimum to a document with supporting material.”
   • Response: They can be found under "02_Response -> 03_Origin_fitting -> Figures_decrease" or "Figures_increase" in the supporting material of the data repository.

1. Comment: “Sections 3.1-3.3: Please consider collected relevant statistical information in a table and remove it from the text to increase readability.”
   • Reponse: We will remove the information from the text and include it in a table.

1. Comment: “L259: change ‘This is correct, as shown by’. Svensson & Ferm did not test this setup and cannot conclude the reason for the underestimation of this system.”
   • Response: We will rephrase this sentence to make it clear that Svensson & Ferm found relationships between air exchange rate and measured NH₃ in their own chamber setup.

1. Comment: “L268-269: Not clear what ‘0.25/a’ means.”
   • Response: 0.25/a is the category of detector tubes from Drägerwerk that can detect between 0.25 and 3 ppm NH₃. We will describe this category in more detail in the revision for clarity.

1. Comment: “L287-311: The effort here to estimate the over- and underestimation is good, but there are some major flaws in my opinion.
   
   Emissions vary over time and as mentioned previously, DTM only provide a spot sample in time. How often should the measurements be conducted to interpolate between measurement points. “
   • Response: Ideally, the DTM should measure emissions at several times during the day to account for diurnal variation. Pacholski et al. (2006) performed them 2-5 times per day, depending on the intensity of the NH₃ fluxes measured.

Comment: “The assumption of zero emission during nights must be justified by literature.”

• Response: We will include nighttime emissions in Table 1, as they are also included in regular measurements by interpolation between the last measurement of the current day and the first measurement of the next day.

Comment: “Why use ‘…’ specifying the ranges?“

• Response: We will remove ‘…’ from the ranges.

1. Comment: “L312-314: I my opinion it can never be recommended to use DTM for absolute emissions, which is also state in some of the literature already cited in this paper.”
   • Response: The literature cited in this paper recommending DTM only for use in qualitative comparisons was referring to the uncalibrated DTM measurements. We will make it clear that we are referring to the calibrated DTM in our discussion and conclusion.

1. Comment: “L320: Kamp et al. found that wind tunnel estimated depended on the air exchange rate of the tunnel as the most important parameter.”
   • Response We will include that the wind tunnel measurements are dependent on the air exchange rate.

1. Comment: “Conclusion:
   
   - It must be clear that there are other alternatives.”
   • Response: As far as we know, there aren't yet any validated alternatives for small plots that don't require lab staff or power. However, we will mention other possible future alternatives in the discussion section.

Comment: “- It should be added that DTM should only be used for comparisons, not determination of absolute emissions.”

• Response: It is not yet clear if the calibrated DTM cannot be used for absolute emissions in all cases. Our experiment has only shown that the detection limit is much higher than originally assumed. This only affects days with NH₃ concentrations below 175 ppb or days with high background concentrations but still below 175 ppb.

Comment: “- Conclusions on wind tunnels should be removed. Wind tunnels were not used in this study, and it is difficult to see on what basis it’s justifiable to conclude that ‘wind tunnels are prone to errors’.”

• Response: Wind tunnels will be removed from the conclusion section.

Comment: “- The reference in the conclusions seems to misplaced as the conclusion should present the most important findings of this work.”

• Response: We will remove the reference from the conclusion.

Comment: “- Include the shortcomings of the DTM method with spot samples in time giving long interpolation periods (unless many manual samples are taken) and the low surface area the chambers.”

• Response: We will mention other possible shortcomings of the DTM in the conclusion after including a paragraph about the shortcomings in the discussion.

1. Comment: “L376: Add README file to the data repository explaining the scripts and data.”
   • Response: There is a readme in the repository that describes what data to find in which folder. The explanation of the R scripts can be found directly in the R scripts themselves.

1. Comment: “Add references to the supporting material (Table A1-4) in the manuscript.”
   • Response: Do you mean that we should cite Tables A1-4 in the Materials and methods? If yes then we will do that.

1. Comment: “Language can be more accurate and clearer. See a few specific suggestions below.”
   • Response: We will make the suggested technical corrections to the text.

Dear referee RC2,

Thank you for your comprehensive review and constructive feedback on our manuscript. We will apply the suggestions you have made as best we can in the revision. Please find our responses to your specific comments below:

1. Comment: “To ensure that readers fully understand the technical details of the instrument setup and what each component is meant to be, I suggest rewording Lines 113-123 and improve the presentation of Figure 1.”
   • Response: We will revise the technical details described in Section 2.1 to make the setup information as clear as possible to the reader. Fig. 1 will be revised to make it self-explanatory, as was also suggested by referee RC1.
2. Comment: “In my experience, calibration of NH3 instruments, even those based on laser spectroscopy, is challenging. The reading of NH3 usually takes long time to stabilize, which largely depends on pump flow rate, but low-frequency drift/variation still exists even under constant room temperature. I’m curious about the calibration of MIRO and Picarro analyzers and their performance during calibration, since it directly influences the results in Fig. 2.”
   • Response: Yes, you are absolutely right, calibration of NH3 instruments based on laser spectroscopy is very challenging. Therefore, the manufacturer of the Picarro NH3 analyzer G2103 used in this study spends great effort in calibrating each delivered instrument already in the factory against a so-called “golden instrument” (S/N: AEDS2079), from which a specific calibration factor for each delivered instrument was derived in the factory. The reliability of this calibration factor was independently confirmed by ab initio calculations using the HITRAN2012 database (Rothman et al. 2013, Journal of Quantitative Spectroscopy and Radiative Transfer 130,4‐50) and by the National Physical Laboratory of the United Kingdom (Martin et al. 2016, Applied Physics B 122, 1‐11). The stability of the Picarro CRDS analyzers was tested in a large intercomparison experiment with 47 CRDS analyzers, and a typical drift of about 0.1% slope/year was found (Yver Kwok et al., Atmos. Meas. Tech. 8, 3867‐3892). On the basis of these data, the manufacturer (Picarro) recommends that “There is no need to perform a true calibration in which the calibration slope is changed according to the results of a direct NH3 calibration experiment.“ Therefore, we chose the Picarro G2103 analyzer as the “master” in our experiment. Detailed information on the traceable calibration of Picarro NH3 analyzers can be found here: https://www.picarro.com/semiconductor/traceable_calibration_of_ammonia_nh3. We will add information on the traceable calibration of the Picarro analyzer to the revised version of the manuscript.

1. Comment: “According to the definitions of detection accuracy and precision (Lines 186-187), Line 205 should be written as “The detection accuracy and precision of the G2103, Drager Tubes and MGA⁷ measurements are displayed in Fig. 2”.”
   • Response: Fig. 2 does not show the detection accuracy and precision of the G2103 as we define it, since the G2103 was used as the reference for the Dräger Tubes and MGA⁷. Instead, we would prefer to rewrite the sentence as follows: “The detection accuracy and precision of the Dräger Tubes and MGA⁷ measurements are displayed in Fig. 2.”
2. Comment: “The authors suggest a more correct detection limit of 152-205 ppb for DTM measurement, but how to explain the significant drop in detection accuracy of DTM from 205 ppb to 305 ppb in Fig. 2?”
   • Response: We are not quite sure why. What was different between the 205 ppb measurement and the 305 ppb measurement was the sampling date. 97, 152 and 205 ppb were measured later in March 2023 to find the detection limit of the Dräger tubes, while the other measurements were taken in October 2022. The temperature in the lab was about 5 °C colder in March, and the batch of Dräger Tubes used could also play a role in the quality of the readings. Another possibility, and probably the most likely, is observer error. It's possible that the Dräger Tubes are not sensitive enough to detect changes in the range of +- 100 ppb. NH3 concentrations are determined by the distance of the discoloration on the detector tubes. It is up to the observer to decide where exactly the discoloration stops. Sometimes the discoloration is only a very light blue, making observation even more difficult. We will discuss this in the revision.
3. Comment: “The detection precision is defined as the relative standard error of all measurements in Line 187, but in the caption of Fig. 2, it is claimed as the standard deviation. Please clarify.”
   • Response: It should be the standard deviation. We will change line 187 accordingly.
4. Comment: “I don’t agree with the statement in Lines 224-225. Fig. 3b clearly shows that the response time of the heated PTFE (4.73 min) is significantly shorter than both PU and Synflex tubing. By the way, check whether the figure 4.73 min is correct, it appears to be 6-7 min according on the y-axis scaling.”
   • Response: While visually it appears to be significantly lower, we used the Kruskal-Wallis rank sum test for significant differences in Figure 3b. This test compares the medians rather than the means, while the figure shows the means. 4.73 min is incorrect, it should be 6.39 min. Thanks for spotting this error. We will correct the text and check for any other inconsistencies between the text and the figures shown.
5. Comment: “Line 281. The expression “A less sensitive detection limit …” is incorrect. It should be “A higher detection limit than originally assumed …” or “A lower measurement sensitivity than originally assumed …”. Similarly, is it correct to say “the highest detection limit …” in Line 285?”
   • Response: We will use "A higher detection limit than originally assumed..." in line 281. We will also change the sentence in line 285 to "Therefore, instruments with high measurement sensitivity are preferred.”
6. Comment: “The main conclusion of this paper is that the DTM is applicable only for large NH3 emission scenario due to its high detection limit of NH3 concentration. Hence, it would be helpful to give an estimate of flux detection limit for the dynamic chamber system with Drager tube used in this study, i.e. the lowest NH3 flux that the system can measure. In Section 4.1, the authors discussed two cases, in which DTM causes flux underestimation and overestimation. I prefer to have a discussion about the flux detection limit in this section.“
   • Response: Our idea was to discuss the flux detection limit at first as well, but to determine fluxes with the calibrated DTM, one would first have to apply the calibration equation from Pacholski, 2006. There the NH3 fluxes depend strongly on the wind speed during the measurement time and are therefore variable. The underestimation of the daily NH3 flux depending on the wind speed would correspond to the daily flux detection limit. We could calculate the detection limit of the uncalibrated DTM. But the uncalibrated DTM is already known to underestimate real fluxes by one order of magnitude (Roelcke, 2002) and should not be used for quantitative measurements anyways. Calculating the flux detection limit of the uncalibrated DTM would in our opinion not add additional value to the discussion. We will instead make it clear that we calculated a flux detection limit range for the calibrated DTM in the table and text.
7. Comment: “Lines 303-304. The NH3 fluxes disagree with those in Table 1.”
   • Response: Thanks for spotting this, you are correct. We initially used the Beaufort scale range 0 to 6 for the information in the text, but later decided to include only the range 0 to 3 in the table because wind speeds above 4 m s-1 would be outside the range of the calibration function from Pacholski, 2006. We will update the text to be consistent with the table.
8. Comment: “Line 305. “daily relative error” means the error relative to the daily emission, but here the intended meaning of the author should be the flux error relative to the total NH3 volatilization for application rate of 60 kg N ha-1 and emission factor of 15%. Please rephrase the relevant content.“
   • Response: You are correct. We will rephrase the sentence in line 305.
9. Comment: “Lines 308-311. The results in the two sentences are obscure, which makes it difficult to understand.”
   • Response: We will rephrase the sentences to make this part clearer.

Response to technical corrections: We will make the suggested technical corrections to the text.