This paper describes two years of eddy covariance CO2 and CH4 flux data from an upland tundra in the Svalbard archipelago. These fluxes are quality controlled and gap filled using standard methodologies. The paper finds that the system is a sink of both CO2 and CH4 but the relevance of this sink to the wider region or global level is not clearly explained. There is also not much discussion of how the magnitude of the sink observed compares to other upland soil measurments in the Arctic.

The authors also perform some correlational analysis of the fluxes against potential driving variables and find that higher CO2 fluxes are correlated with high wind speeds and high temperature anomalies. The CH4 fluxes in contrast are not found to correlate with expected driving variables like soil temperature and the authors propose that a mismatch in the depth of the microbial community and the soil temperature may explain this but do not explore if they can recover the expected correlation with lag analysis. The authors also describe a seasonal cycle in both fluxes with a maximal sink in the summer and near 0 fluxes in the winter. Other relationships (including null relations) with potential driving variables are not reported and the diel cycles in the flux data are not discussed. 

The authors also propose explanations for the flux correlations and seasonality but they do not seem to be strongly supported by other lines of evidence. Additionally, although the authors discuss that the fluxes and changes they observe may be the result of microbial respiration or CO2 uptake by plants they also do not report partitioned CO2 fluxes which would allow them to more directly attribute the correlations and seasonal changes they see.

Additionally, open path CO2 sensors (including the LI7500a) have been reported to show an artifact that results in a negative correlation between H and CO2 flux during cold conditions (https://doi.org/10.1175/JTECH-D-17-0085.1). The correlation observed by the authors may be related to this rather than a real biogeochemical effect and they should discuss how the presence of this artifact may affect their seasonal interpretation and if using the correction term proposed in the above paper affects their results. I also am unsure if the difference in how the authors treat the friction velocity threshold filter for their CO2 and CH4 fluxes is appropriate.

I think that the data reported in this paper are of interest to the arctic biogeochemistry and eddy covariance community. However, I think that the analysis of the data and the discussion of its significance could be refined and expanded and that doing so would improve the manuscript and its signifcance.

37: I feel like this sentence repeats information given in 34 and that the information given in the first paragraph could flow better

45: Given your focus on Arctic soils, is this reference relevant

54: This isn’t very clear and your sentences can be restructured to flow better, in the previous sentence you mention the permafrost becoming a source because of more accessible OM and then mention that the driver of the net uptake is vegetation? It might be helpful to clarify how despite there being net uptake, permafrost thaw allows for more emissions

68: Is this effect referring to the effect in wetlands?

112: I would change the phrasing to “The measurement campaign ran from”

194: In this case does your sensor report CO2 uptake during the winter months? How do you deal with this?

211: Since the friction velocity threshold represents turbulence being underdeveloped, when it is below the threshold, wouldn’t all the fluxes including the CH4 fluxes be incorrect as they are also affected by the low turbulence?

249: Since you discuss results from other seasons, I think it would be good to mention the footprint coverage in them as well and if they differ from the summer? If they are different, it may complicate your interpretation of seasonality.

255: It is a bit unclear to me how the end of the thaw and freeze up periods are defined. The text only mentions the start.

273: The definite article “the” seems to be missing at the start of this sentence and in a few other places

309: Is the paper you reference in comparison in a similar ecosystem and location?

318: Since the CO2 fluxes are used to gap fill the CH4 fluxes, how can you be sure this similarity is not a result of that?

324: I’m not sure if reporting the mass budget without gap filling is meaningful. If the gaps are not randomly distributed then the periods where gaps are less common will be overrepresented in the budget if you don’t gap fill so it’s not as directly relevant to the ecosystem fluxes.

352: It has previously been found that there may be an artifact that causes a correlation between H and CO2 flux in open path sensors during the cold season so the correlation you find may also be related to that. WIthout further discussion or analysis of this effect,  I wonder if it is justified to say that this is strictly a real effect?

365: I wonder if a lot or most of the correlation you see with wind speed is just due to the correlation with wind speed? It seems like most of the data on the at high wind speed are also at low H.

366: Which methane concentration gradient are you referring to here? I don’t recall you mentioning soil methane measurements or a vertical profile in the atmosphere in your methods?

369: I don’t follow how the concentration gradient implies aeration rather than methanotrophic rates.

379: In a section further up, you mention that the soil temperature in the summer (line 321) is the driver of the increased methane uptake in that season. How do you reconcile this with the lack of correlation here?

383: I assume that soil temperature in the near surface layers is lagged (ie the soil near the surface heats up before the deeper layers) If there is a question of the soil depth of the methanotroph community, would a time lagged version of the 10cm soil temperature data show a better correlation with the methane fluxes?

386: I’m not sure what psychrophilic means coming from an eddy flux background. It may be helpful to define this?

390: Are these anomalies over the mean annual temperature in this period or over the day of year averages? It is not clear from the text what these are and I think you should clarify them.

400: Can you quantitatively assess the strength of these trends ? It is hard to tell how significant some of the trends are especially as some seasons have small variations in the x axis? Additionally, does this mean that on days where the temperature on a given day is above the long term seasonal mean for that day the fluxes are higher?

420: Regarding CO2 in terrestrial systems, the flux is often partitioned into productivity and respiration. It might be interesting and clarify what drives changes you observe to consider the factors driving each of these separately.

424: I’m not sure how you measured soil aeration

427: It is not clear to me why you would see a negative CO2 flux in the winter. Is this flux below your measurement error and detection limit?

431:Is it correct to say this since you find only a very small relation with soil temperature?

455: Is it a typo that the summer shortwave radiation is lower than the winter one?

Review Donateo et al.

General comments

The dataset presented in this manuscript is of high quality and is a unique time series of net CO2 and CH4 fluxes from an understudied ecosystem type in the Arctic, that none the less comprise the majority of the Arctic area. Thus, understanding the magnitude of CO2 and CH4 exchange with the atmosphere is relevant in the context of the Arctic carbon budget and how it responds to environmental change. 

The authors present and discuss the data in a concise and structured manner with good interpretation, although I recommend more careful interpretation of the relation between wind and net CH4 uptake by including the interaction with soil hydrology. Also, the authors do not show this correlation, which I suspect is very variable and not better than the one for temperature.

There is a general switching between present and past tense in the entire manuscript which makes it confusing at times to read and halts the flow. Language itself is good but I strongly recommend you to make it either present or past tense throughout.

Perhaps I am missing a little more discussion on the annual sums of CO2 fluxes compared to other studies in the Arctic.

I also suggest to add an analysis of the diurnal patterns for the different seasons as this would support the overall purpose of the manuscript.

Abstract

Line 25-27: This formulation “CH4 fluxes…” is not very clear, because how is the reader supposed to understand how wind velocity is related to electron acceptors in the soil. There seems to be missing some information in between the wind and the soil, so please detail this

Line 27-28: “High temperature anomalies…” Do they occur in winter/autumn/spring? Please clarify in text. Also, how does this respiration inhibit productivity? Do you mean that these anomalies decrease the annual net sink?

Line 28-29: How much of the annual budget does the winter CH4 emission constitute? I think you should add that number here

Line 29: replace “comprehending” with “understanding” or “measuring”

Line 30: replace “elucidate” with “constrain”

Introduction

Line 37: delete “affect the…”

Line 47: “absorption” Do you mean “net uptake”? If so I suggest to write this. Merge this sentence with the next by deleting “This is” and replace “that season” with “growing season”.

Line 51: I would here use “soil hydrology” instead of “water levels”

Line 53: I do not think the Kleber et al. 2023 citation fits in here as this relates to glacier retreat and release of thermogenic methane. So I suggest to remove it

Line 55-64: I suggest to delete this entire paragraph. It is kind of trivial for the interested reader for your paper. No need to use space on this and in the next paragraph you get to the primary knowledge gap which is the dry tundra ecosystems

Line 76-90: Overall a good wrap up of existing papers on net CH4 uptake in dry tundra soils. A few updated papers came out recently that might be of interest for you as well:

• Juncher Jørgensen, C., Schlaikjær Mariager, T., & Riis Christiansen, J. (2024). Spatial variation of net methane uptake in Arctic and subarctic drylands of Canada and Greenland. Geoderma, 443, 116815. https://doi.org/10.1016/j.geoderma.2024.116815
• D’Imperio, L., Li, B.-B., Tiedje, J. M., Oh, Y., Christiansen, J. R., Kepfer-Rojas, S., Westergaard-Nielsen, A., Brandt, K. K., Holm, P. E., Wang, P., Ambus, P., & Elberling, B. (2023). Spatial controls of methane uptake in upland soils across climatic and geological regions in Greenland. Communications Earth & Environment, 4(1), 461. https://doi.org/10.1038/s43247-023-01143-3

Line 96: I do not understand the phrase “equilibrium between CO2 and CH4 in dry tundra environments”. Do you mean the balance between the net flux of the two gases? I suggest to rephrase to make it clear what you mean

Line 97: You have to mention in the very top of the introduction that one of the predicted changes with the Arctic Amplification is “increasing frequency…”. In this way you tie the introduction better together

Line 100: This sentence is a little off in relation to the text above and below, so I suggest to delete

Line 103: Write how many years

Line 105-107: When you write “assess impact on…greenhouse gas fluxes” do you then mean the mechanisms in the soil or the impact on the budget? I think it is important to distinguish as most of your introduction is focused on budgets, which essentially your study design by using eddy is best suited for. So for the reader it is important that you upfront outline whether you focus on mechanistic processes or budgets.

Methods

Measurement site

Line 137-140: I guess you mentioning this as it might be affecting measurements occasionally? I think you just have write why you mention this.

Eddy covariance data analysis

Line 224: You develop a random forest model to account for the complex biogeochemical variables using 12 drivers, but only one is directly related to the soil. Thus, your random forest model is less of a direct model but rather an indirect based on meteorological variables. I would therefore be hesitant to call this a model that can handle complex biogeochemical interactions as you state.

1. Results and discussion

Line 251 – 270: I think all this text belongs in Methods section, perhaps as a sub chapter to the eddy covariance data analysis section.  

3.1 CO2 and CH4 mixing ratio

Line 300-301: The Wang et al. 2013 is hardly referring to High Arctic conditions and I would not cite this paper here and suggest to use a more suitable one that represents Arctic conditions

Line 307: replace “absorption” with “net uptake”. Absorption is a word used in chemistry

Line 297 – 309: In this text you mostly present the median values, but I think you neglect to address the sizeable variation of instantaneous fluxes around this median for both CO2 and CH4. Your box plots off course show that most of the fluxes are within the 25^th and 75^th percentiles, but regardless of season you have very high variation of fluxes within very short times. For CH4, for example, you fluxes range from +10  to -10 nmol m-2 s-1, in what seems to be hours or even shorter (difficult to see on figure 3). Hence, what is behind this large variation in flux estimates. IS it purely stochastic related to the measurement principle due to the high noise-to-signal ratio for CH4, at least? I doubt that it can be related to a process that can switch the flux that fast from strong net uptake to strong net emission. Please comment on this in the text.

Also, I missing an analysis of the diurnal variation in CO2 and CH4 fluxes for the different seasons. One would expect higher variability in summer than in winter. With your data set you are able to make some robust estimates of diurnality. This also feeds in to your overall aim of investigating temporal patterns and drivers of CO2 and CH4 fluxes. So I suggest you include this in the manuscript. If there is no clear patterns then you can choose to leave the data out of the paper and merely mention that no differences in diurnal patterns were detected and leave it at that. However, if there are some differences it might serve the overall purpose of your manuscript very well.

3.2 CO2 and CH4 mass budget

Line 332-333: Unclear what number you refer to in Treat et al. 2016. Please specify whether it is CH4. Also there is no comparison to other Arctic CO2 budgets although there have been measured a lot, so I suggest you compare the CO2 budgets with other studies.

3.3 Physical drivers on GHGs surface fluxes

Line 357: I think you refer to Fig. 5b and not “9b”

Line 372-375: I think the discussion on methanotrophy in Type I and implicitly Type II is off here and not really relevant as you have to assume that it is aerobic methanotrophy that is responsible for the net uptake, but whether it is one type of methanotrophs or the other is irrelevant in your case as you cannot really evaluate this. If you were to say anything then you would have to assume that it is type II (high affinity MOB’s) that do the oxidation as there are no wetlands in the CCT footprint.

Line 375-377: This is more a conclusive statement and I would remove from here and add to conclusions if needed

Line 386-388: But on the other hand windspeed is higher in winter with the lowest fluxes. And given the low range of wind speeds I doubt that you will find a strong link between net CH4 uptake rates and wind speed. You are correct that oxygen addition to deeper layers is likely stimulating methane oxidation, but this is as much related to the drying out of the soil during the summer. Since you say the CCT is also a semi-desert my best interpretation is that the CH4 flux regulation is most directly related to the soil hydrology, indirectly affected by the wind that can dry the soil, rather than it is actually mechanical mixing of oxygen in to the soil. Remember also, that methane oxidation requires one O2 molecule. Thus, the oxygen requirement is very low compared to how much O2 there is in the air, it is several orders of magnitude. So I would also assume that even at atmospheric stability or inversion and when the soil diffusion was not restricted by water (summer and freezing) O2 supply would not be limiting CH4 oxidation. I therefore have difficulty in attributing turbulent vertical atmospheric O2 mixing as the primary stimulant of soil profile CH4 oxidation. Research points to that this is regulated by soil hydrology and porosity. I therefore suggest you to moderate this claim and include how wind may act to dry the soil and hence increase diffusivity for O2 and atmospheric CH4. However, the problem is that you do not have direct soil moisture measurements to support this claim, but still I think it is a more feasible explanation that has been shown in other studies.

3.4 GHGs fluxes response to seasonal temperature anomalies

Line 400 Do you mean net annual CO2 uptake?

Line 401-402: This statement would be easier to interpret if the temp-anomaly vs CO2 plot were split in to seasons. Perhaps do this as a supplementary figure?

Line 411-412: Difficult formulation. I do not understand what you mean. I suggest to rephrase. Is this transition in the same direction in summer or winter or in opposite directions and hence what is the impact of the annual net CO2 sink? If the transition to higher net CO2 emissions with temp anomaly is the same for summer and winter then the net annual CO2 uptake decrease, but if the transition is opposite in winter and summer then they may cancel out. I think this is what you mean by the sentence, but I am not sure. Sorry for the confusion.

1. Conclusions

Line 424: “increasing aeration” In my opinion and from knowledge of these soil types, they dry out fast in the spring if not snow covered. This is maybe because of wind, but likely also because of a relatively low albedo

Line 428: “reducing any further biological activity.” I suggest to delete this part of the sentence as it does not really add meaning to this conclusion

Line 431-433: A weakness here of the conclusion that “other environmental variables” control CH4 uptake is that you have not measured soil moisture content which is at the core of this increase. Solar radiation and wind plays a role in the speed of drying, but the soil material and structure ultimately determines how much it dries under the given climatic conclusion. I suggest you attempt to include more reflections on the role of soil moisture and that you may have a missing link here in your study.

Line 433-434: I do not think this is the case. For example, Jørgensen et al. 2015 that you also cite shows that methanotrophs in dry tundra has a Q10 of 2, which does not indicate a lesser temperature dependency. Rather it is likely that the variation in CH4 uptake is not limited by temperature, but by other factors

Line 434-435: I think you have an indirect effect of the wind, but via the soil hydrology as mentioned above. Furthermore, this correlation is not presented and the reader cannot assess if it a strong or as weak a correlation with the soil temperature. So I would be careful in concluding like this here and rather moderate the discussion in throughout the text.

Line 438: Maybe add here after “…CO2” “both in summer and winter periods, effectively reducing the net annual uptake.” I think this is an important finding.

Line 438-439: Rather it is the opposite. Higher temperatures would stimulate plant growth if not limited by water and hence higher GPP. But if this higher GPP is counterbalanced or even exceeded by more frequent temp-anomaly driven CO2 emissions in summer and winter the annual net effect may actually be an overall decrease. The way you write it here indicates that CO2 respiration from increasing temp inhibits plant productivity, but this is not the case.  So rephrase to avoid this mistake.

Figures

Really nice figures 1 – 6

Figure 4 – lower panel for CH4. Check y-axis title. There seems to be an “m” too much

Figure 6 – In the caption the black line is not explained. Also, I would suggest to show this figure split in to the different seasons, so in order to more clearly see if seasons behave similarly or different. From the text