This study focuses on intra- and interannual trends of GHG emissions from 4 small urban ponds and also addresses their underlying drivers. I believe that the study is a great addition to inland water studies, especially since long-term data on small pond GHG emissions is currently scarce. I commend the authors for curating such a comprehensive dataset. However, the current structure of the paper needs to be revised to make it easier for the reader to follow and also for possible final publications to Biogeosciences. Below are my brief and detailed comments for each subsection for possible improvement of the manuscript.

Brief comments

Abstract and introduction

While the introduction was well compiled with the key motivation of the study, the end part did not include clear objectives and hypotheses for the work, which would have guided the reader better throughout the whole manuscript. I suggest that the authors include this and also adjust the intro to highly key gaps that would be addressed later on in the manuscript

Materials and methods

Although the description of the analysis of the GHG and associated parameters was well done, the statistical analysis part was too short and lacked enough detail. For example, the authors said they used a one-way ANOVA, yet they had a two-factor problem. i.e., seasonal and also pond-type influences. Also, what post hoc tests were used, what correlative analyses were used, and the main aims of this analysis are either lacking or not clearly stated. I suggest that more details should be added addressing the points above.

Results and discussion

While combining the results and discussion is an acceptable practice, it sometimes leads to sectors of the manuscript that are not well described. For example, the trends of CO2 and N2O were only mentioned as results and not well substantiated by findings from other studies. In contrast, CH4 trends were well described by the authors, with proper discussions also related to other studies. I suggest either sticking to methane alone or giving also equal focus to the other GHGs. The manuscript also has 13 figures. While this is fine, readers may end up missing the most crucial part of the results. The unwritten rule of thumb is 6 to a maximum of 8 graphics, which include both tables and figures. I suggest the authors reevaluate the key figures guided by the objectives of the study and then reduce the current number and keep the rest in the suplimentary. Most of the results also lacked tests of significance and I suggest that this should be included in the revised drafts. If differences are not significant, its always acceptable to refer to them as trends

Conclusion

It needs to be focused on the objectives of the study and also to have a general outlook on the potential of urban ponds of inland water GHG dynamics.

Detailed comments

Abstract and introduction

Line 15,16: Consider mentioning the direction of the relationship.

Line 20: Consider adding a statement relating light availability in the clear ponds to enhance macrophyte growth. The current statement may be unclear at first read.

Line 21: Trim down the statement about pond methane fluxes, for instance, 'Pond methane fluxes to the atmosphere… '.

Line 37: Consider rephrasing: Greenhouse gas emissions from inland waters to the atmosphere….

Line 39: GHG emissions from lakes….

Line 44: Remove the …. You can replace it with such as, which indicates that these are just examples and there could be more.

Line 46: Noun required after the word this…for example, this finding, this conclusion….check here and everywhere in the manuscript.

Line 52-54. The sentence on runoff comes from nowhere. Did you mean the rainfall runoff gets into the ponds? Consider revising it to make it clearer.

Materials and methods

Line 99. Did you mean the institute laboratory? The use of a home may imply a laboratory located in a personal house/apartment.

Line 109. Consider revising from “consistent in” to “consisted of”

Line 111. The statement is a bit confusing. Consider revising it to make it clearer. How were the gases measured with 60ml syringes?

Line 113. Consider revising the statement. Did you mean that the measurements at Silex were of a longer frequency?

Line 210-211. How were seasonality and pond type considered in your ANOVA analysis? The current statement is too short and lacks details. Was the ANOVA a repeated measures ANOVA as you sampled on the same pond multiple times?

Results and discussion

Line 223. Wetter and colder

Line 228-229. Consider adding the reference period at first mention and not at the end of the statement

Line 233. Missing article; “with the silex pond”

Line 244. In Figure 3, I suggest adding letters to the boxplots to indicate significant differences from the ANOVA test. This will help the reader quickly follow the graphs and avoid looking at an extra table in the supplementary.

Line 252. Are you reporting significant differences or trends? Check here and everywhere where you report comparisons of means. Also, indicate the level of significance as the information is currently missing in the results

Line 253. I would move the explanations to the discussion, i.e., owing to primary production…

Line 256. Consider using low instead of minimal

Line 257. Same comment as 253

Line 258. Replace Maximal to High

Line 259-263. Correlation results are important for GHG process information as also discussed in this paragraph. I suggest moving them to the main text and, if possible, using scatterplots for the main relationships and indicating the correlation coefficients in the graphs. Also, always include the direction of the relationship, i.e., how was pco2 related to precipitation? Was it a negative or positive correlation?

Line 264-266. I now see that the results and discussion are combined. While this is fine, the way it's currently written includes a lot of speculative statements that have not been substantiated by the findings of other studies. I suggest taking a closer look at all statements made and trying to support them with other studies. Putting a citation at the end without stating where the authors found similar results is also not encouraged. You can use (e.g., ……) in the citation to make clear that these authors found similar findings.

Line 269. See comment on line 253

Line 277. See the comment on the use of “this” above

Line 278. Add a comma between ponds and the

Line 279. Were these differences based on the other factors also tested, i.e., the effect of the size of the pond? This analysis would validate the statement. At the moment, it’s a bit speculative

Line 282. Citation of figure or table needed here.

Line 284. Were surprising…

Line 291-295. This paragraph is a bit confusing. I know what the authors mean, but I suggest it be rephrased in order to explain better the lack of correlation between N2O and DIN and its link to nitrogen deposition. How much is the nitrogen deposition in the region and how does it decrease from the edges of the city to the inner parts? Without this data, the current statement is somehow speculative

Line 297 -298. How do these bubble fluxes compare with other values from similar studies? Are they on the higher end or lower end? I suggest adding a few comparison studies in all fluxes reported to give an idea of where your study stands in terms of the magnitude of the fluxes.

Line 304. I suggest adding the equation of the fit on the graph.

310-312. This is what I mean by referencing of other studies to support your findings/

Line 337. I suggest always using, e.g.,…. Or “similar to what was found…” for every citation quoted in the discussion, particularly those that involve specific findings. This form of citing guides the reader better.

Line 338. I suggest adding the equations of the relationships to Figure 6, which may be useful for future comparisons with other studies and also allow them to be potentially used to estimate ebullition methane fluxes where temperature data is available, as this study has done.

Line 356. Add letters from posthoc tests to indicate seasonal differences to this figure, similar to my comment on Figure 3

Line 373. Than for diffusive fluxes…

Line 389. How do you explain the polynomial U fit in the first panel?

Line 393. Was there a statistical test to show that the differences were significant? Judging by the error bars, which sometimes overlap, it may be that the differences were not significant, but I do agree that the trends are there. In cases where the relationships are not significant, I suggest sticking to … showed trends of being higher in…even though the difference was not significant.

Line 401. I now see the explanation for the polynomial fit, which also makes sense. However, this may not be so clear at first read. Hence, it may help to reference the result first and then link it to the explanation. Also, has the relationship with phytoplankton been found in other turbid pond studies. The current reference talks about lakes.

Line 411. Where is the regression done in the results?

Line 421-423. Lakes and ponds are used as synonyms here, even thou they may have different characteristics. Check here and everywhere to ensure that references made on lakes are assumed to be related also to ponds.

Line 472. “as” not “than”

Line 485. Modify the graph to include posthoc tests showing significance across ponds and seasons

Conclusion

Line 557-562. Reads more like the results and discussion part. I suggest rewriting the conclusion part to focus more on what were the objectives, what were the main conclusions from each objective, and finally future perspectives on what can be done better.

General comments

Authors assessed various aspects of GHG dynamics in four urban ponds of either macrophyte or phytoplankton dominated stable states over a 2.5 year period. The authors have produced an impressive and valuable long-term dataset on greenhouse gas (GHG) dynamics in ponds, which notably includes the often-overlooked ebullitive flux and provides insight into various methane pathways. While the data and results are impactful, the manuscript requires major revisions to be considered for publication. First, the writing needs major improvement for clarity and quality, and second, the authors provide insufficient details about methods and statistics, and should reconsider how their gases are presented. Below are my general comments for each section, followed by more specific line comments. I believe these data can make a good contribution to the body of literature on GHG dynamics in urban ponds, and hope these comments will help improve the manuscript.

Abstract

The abstract was long and wordy with the results. It may be helpful to be more concise and summarize some of these major findings. I provide examples in specific comments below.

Introduction

The introduction has some good elements to it but overall lacks supporting information for much of the content covered in the study. For example, authors compare macrophyte versus phytoplankton dominated systems but only provide background information on the impact of macrophytes to GHGs. One of the most interesting components of the study to me is methanogenic pathways and methane oxidation, which has not been covered in urban ponds to my knowledge. While authors cover GHG fluxes and drivers, no mention of methane oxidation, methanogenic pathways, and their significance are made in the intro and is only briefly mentioned in the concluding paragraph. I’d like to see some supporting information for phytoplankton dominated ponds, methane oxidation, and methanogenic pathways in the intro.

In the closing paragraph the authors do not include any objectives or predictions/hypotheses, but rather focus on some of the methods. I strongly suggest focusing less on methods and including objectives and predictions/hypotheses to help guide readers.

Methods

The methods section requires some reorganization and lacks a lot of details.

The statistics section is grossly lacking in detail and what methods were used appear concerning, but potentially due to no explanation of the approaches used. Authors need to explicitly state when/why they use one-way ANOVAs (we use one-way ANOVAs to test for the effect of X on Y and the effect of A on B) and linear or exponential regression. Exponential regressions were used but no mention was made in methods. I also don’t think that one-way ANOVA is appropriate where it is used. First, if your analyses are including all data over time, and there are repeated observations from the same four sites over time, that is a case of pseudoreplication. I suggest looking into generalized linear mixed effects models (GLMM) to account for time as a repeated measure.Second, as an example looking at Table S6 (a bit hard to interpret), I think what I’m seeing are pairwise comparisons for one-way ANOVAs that looked at the effect of pond and season on a variable listed in a column? (i.e., chl-a ~ pond + season)? If so, I think these are instead two-way ANOVAs, and the type of pairwise comparison needs to be stated. Table S3 suggest PERMANOVA was used but again, this is not clear. Stats for pairwise comparisons are provided but nothing for the model itself. Degrees of freedom would be a helpful term to report along with the model stats, not just pairwise stats.

On another note, I can understand using ANOVA to see significant differences between sites for water chemistry variables (chl-a, TSM, %O2), but later on (e.g., Figure 8, Figure 10, Figure 12) linear regressions are used to test for effects of environmental variables on GHGs, which is redundant. While I don’t think it is wrong to use multiple individual regressions to test for the effect of each environmental variable on a gas, multiple regression models may be more informative, and again, you can account for the pseudoreplication of repeated measures over time. Last suggestion here, the point of the paper is to look at differences between macrophyte versus phytoplankton dominated ponds. Have authors tried grouping by stable state type (macrophyte or phytoplankton), and testing for the effect of that? Two sites per level might not be sufficient enough but curious if this was considered. If you went this route and used a GLMM (for example), perhaps individual site could be set as the random effect.

Further, I have concerns for how gas concentrations/quantities are presented and equilibrium saturation is calculated. Why present CO2 as a partial pressure, methane as a concentration, and N2O as a percent saturation compared to equilibrium?? I strongly advice all three gases be reported in comparable molar units. In addition to molar concentration, authors should report their deviation from equilibrium, either as a % (like N2O.  but values will likely be too high for CH4), or the factor of super/under saturation, and be consistent for all gases (i.e., the concentration of CO2 was X umol/L and 12-fold supersaturated compared to equilibrium). Deviations from equilibrium are more informative for biological changes to gas concentrations. Alternatively, there is so much information in this manuscript that authors should consider removing results for gas concentrations altogether and focus on air-water fluxes, as some seasonal patterns and environmental drivers appear somewhat similar between concentrations and gases.

Results/discussion

This sounds more like a results section than a combined results and discussion section. I recommend keeping results and discussion separate. I provide specific comments up to some of the results, as I imagine results reporting will change when statistics are improved, and discussion will change when these sections are split apart.

Conclusion

Major results should be broadly summarized here but the significance of the work should also be included. If authors include predictions, they can be circled back on here as well.

Specific comments

ABSTRACT

Line 8-9: Suggest rewording as “…but it is unclear if these two states affect the emission of greenhouse gases carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O) to the atmosphere.”

Line 9-12: Suggest rewording to something like the following and including “fluxes”: “We measured the saturation and air-water flux of CO2, CH4, and N2O gases, and ancillary variables 46 times over 2.5 years in four urban ponds in Brussels, Belgium: two clear-water macrophyte dominated ponds and two turbid-water phytoplankton dominated ponds.”

Line 12-15: Here and throughout, I suggest authors use first person instead of passive voice. I reword the next two sentences to include the objective up front and change to passive voice. I also include the method for ebullitive fluxes: “To quantify CH4 ebullitive fluxes we conducted 8 bubble trap deployments totaling 48 cumulated measurements. To characterize methanogenic pathways (acetoclastic or hydrogenotrophic) and quantify water column methane oxidation (MOX) we measured the 13C/12C isotope ratio of CH4 (δ13 13 C-CH4) from bubble traps and sediment bubbles.”

Line 15-18: These results could be removed from the abstract. Temperature and precipitation are already touched on later when discussing fluxes.

Line 18: Remove “The sampled”.

Line 22-23: The sentence beginning with “The temperature sensitivity..” could be removed or combined with previous.

Line 23-28: These sentences could be combined to reduce wordiness.

Line 35: I suggest adding a concluding sentence to highlight the implications and usefulness of the results.

INTRODUCTION

Line 39-41: This sentence leads me think you are going to further discuss lentic versus lotic GHGs. I suggest replacing with a sentence highlighting estimated GHG emissions from lakes combined to guide the reader into the next sentence about small pond contributions.

Line 42: change “could be” to “are” or “can be”

Line 42: shallow lakes are not always ponds (see Richardson et al. (2023) on defining ponds, DOI: 10.1038/s41598-022-14569-0). Maybe cite Downing (2010; DOI: 10.23818/limn.29.02) to make the point in this sentence for small ponds?

 Line 44-49: Not sure if I would say artificial ponds are “seldom” investigated these days, maybe that the body of literature is growing. Other artificial and/or stormwater pond papers looking at GHGs and carbon inputs: Goeckner et al. 2022 (DOI: 10.1038/s43247-022-00384-y), Ray and Holgerson 2023 (DOI: 10.1029/2023GL104235), and Kalev et al. 2020 for DOC and POC inputs (DOI: 10.1016/j.scitotenv.2020.141773).

Line 50: I’m not sure that I agree that urban ponds are mostly in green spaces. If you are referring to a particular region, I would specify that, but this point contradicts what you say in the next sentence that they are surrounded by impervious surfaces.

Line 53-54: This sentence is redundant with the sentence on lines 46-48 on C & N inputs. I suggest moving this up to replace that sentence and added a concluding sentence here that highlights a knowledge gap covered in your study.

Line 55-56: This sentence is a little confusing to me. When you say submerged aquatic primary production, are you referring to the contribution of submerged aquatic vegetation to primary production? If so, I don’t think phytoplankton is typically referred to as submerged vegetation. I would simply say primary production or reorganize the beginning of this paragraph to begin with the alternative stable states.

Line 57: Indicate which stable state is associated to clear or turbid water.

Line 58: Macrophytes also impact CO2 cycling (e.g., in the Theus et al. 2023 you cite in the next sentence). Further, no background is provided for the effect of stable states on CO2 at all in the introduction. This should be included as it is for CH4 and N2O.  

Line 62: Ojala et al. 2011 may also be a relevant paper to check out but they focus on clear versus brown-water lakes (DOI: 10.4319/lo.2011.56.01.0061).

Line 67-70: I would combine these sentences to highlight where positive N2O-macrophyte relationships have been reported and save the details for the discussion.

Line 71: Authors haven’t described why denitrification and N2O are associated. I would either include their association (i.e., that N2O can be produced is an intermediate product of denitrification or nitrification), or remove and save this for the discussion.

Line 72: I suggest concluding with a sentence on the significance of quantifying GHG fluxes in macrophyte vs. phytoplankton dominated systems. Further, no background information is provided on GHG dynamics from phytoplankton dominated ponds, of which there is plenty of literature on.

Line 73 / paragraph 3: This paragraph is good. I suggest adding some support from Ray and Holgerson (2023) on the contribution of ebullition to CH4 fluxes in artificial ponds. Also, you don’t mention anything about methanogenesis or methane oxidation until the closing paragraph, whereas this is a profound and really interesting part of your work! The significance of understanding methanogenic pathways, and methane oxidation, should be included in this paragraph.

Line 81 / paragraph 4: This whole paragraph should be re-written to outline the objectives of this study. To me they were (1) to understand annual variability in saturation/fluxes, (2) characterize and quantify CH4 cycling pathways (methanogenesis/methonotrophy), and (3) identify drivers of these fluxes/pathways including pond type and environmental variables. Then you can briefly say you collect 2.5 years worth of data (so impressive!) on GHG dynamics and environmental conditions in four urban ponds of differing stable states. I also suggest adding predictions based on supporting information provided earlier in the intro. Then conclude with why the study contributes to the body of lit on urban pond GHG dynamics.

METHODS

Line 98: Did you visit each pond on the say day? What time of day (approximate window) did you sample? If you remove the methodological specifics from the conclusion of the introduction, add the details here about the 46 sampling days and period of time you sampled sites from (June 2021 – Dec. 2023). Also, how would you describe the climate/precipitation of Brussels? These can be added before the sentence here.

Line 98-99. I would separate GHGs and “other variables” here and focus on the “other variables” first. Then move on to GHG sample collection. In any case, how far below the surface did you collect water from?  

Line 100-103: pCO2 analytical approach needs to be moved to the same section as CH4 and N2O unless it was a portable analyzer (unclear). When you say headspace approach, are you referring to the headspace equilibrium approach? If you used the same approach following the cited Borges et al. (2019) then I think so? Unclear. If so, more information is needed here for the headspace approach. How much water volume versus headspace volume did you equilibrate? How long did you equilibrate? Did you use N2 gas as the headspace or ambient air??  

Line 102: After each cruise? Do you mean when you sampled a pond from the pontoon? I suggest saying “before and after each sampling event”. Also, is the Li-Cor Li-840 a portable gas analyzer? I didn’t think so but now I’m wondering if it is. If so, measurements of CO2 in the field needs to be explicitly stated and the approach described better.

Line 104: Change “in” to “into” and “poisoned” to “preserved”.

Line 117: How did you store the gas prior to analysis? Same with other types of samples collected, storage prior to analysis should be included.

Line 125: Ok so for CH4 and N2O you collected water, then used the headspace equilibration approach in the lab? If you also used this approach for CO2 but it was done in the field for CO2, this still should be described earlier.

Line 130-131: I strongly advice authors to report each gas as both a concentration and some form of their deviation from equilibrium. Authors say reporting pCO2, nmol/L of CH4, and %N2O is “with convention in existing topical literature”, but no references are provided, and I disagree with this approach. Other impactful pond GHG papers focusing on concentrations alone maintain the same units (e.g., Holgerson 2015, DOI: 10.1007/s10533-015-0099-y), and this allows for easier comparison. If N2O is presented as a deviation from equilibrium, I think the same should be included for CO2 and CH4, as deviation from equilibrium is insightful for biological changes to these gases. This helps readers understand to what degree the gases are “systematically and distinctly above saturation”.

Line 133: How did you calculate the equilibrium solubility of N2O in water?? Did you calculate N2O solubility using the water temperature at the time you collected samples (based on Henrys law)?

Line 142: I suggest moving this section above the GHG analytical section to improve the flow of methods. (GHG analysis -> GHG calculation)

Line 154: Is this DIN the sum of NH4-N, NO3-N, and NO2-N? or the sum of the full concentration of each?

Line 158: Finally I see that CO2 was measured in the field with the Li-Cor Li-840. This needs to be made clear much early on...

Line 160: Change “on” to “in”.

Line 160: Where did you get this value of 1.9 ppm for CH4?

Line 160-163: I’m a little confused. Because authors use the phrasing “equilibrium with atmosphere for N2O” it sounds like equilibrium solubility in water. Because authors also say air mixing ratios here, I think that this is not what authors intend and instead they mean the atmospheric concentration of N2O.

Line 211 / statistics: This statistics section requires much more detail and I’m a little concerned about the statistics overall but this may be because things are hard to follow. I cover my concerns in the general comment above for this section.

RESULTS & DISCUSSION

Line 237-238: this sentence is redundant with the previous where you already report these values. I would remove and site Figure 3 with the previous sentence.  

Line 252: Were these in the surface of the water? Would be helpful if the depth of measurements are specific in methods.

Line 256: The range is from 40 to 13804 ppm but in the methods authors stated that pCO2 was “systematically and distinctly above saturation level”  which was cited at 400 ppm. I’m curious now if authors meant atmospheric CO2, of CO2 dissolved in water, which would differ based on temperature.

Line 257: warmer waters also hold less gas. Including percent or factor of saturation compared to equilibrium may be helpful to understand lower CO2 concentrations in the summer.

Line 261: What kind of model was used for the results in Table 3? Are these results from a PERMANOVA? If so, nothing about a PERMANOVA was included in the stats section. Again, pseudoreplication from repeated sampling over time. There are also linear regressions of these gases over some of these variables in other analyses, which is redundant.

Line 270: What do you mean by “sometimes correlated”. As in during certain seasons or only in some ponds?

Line 283: Not having explanations for correlations included in models is a good example of why a prior hypotheses are useful. There is already a lot of information in this manuscript so maybe some of these analyses for drivers of gas concentrations are not needed in the first place? Just a thought.

Line 286: This makes me wonder if this DIN is the N fraction of inorganic nitrogen forms or their full concentration (i.e., NH4-N versus NH4) and what the difference would be for results here between either option, but I think the sum of N fraction in the inorganic forms is more appropriate.  

Line 295: Would be helpful to summarize at the end of this section what the differences in GHGs for macrophyte versus phytoplankton dominated ponds.

 Line 348-349: The observation of higher CH4 in summer and spring was already noted for CH4 concentration. Maybe a good example of why reporting concentrations and fluxes is not needed?

Line 367: See also Ray and Hoglerson (2023), DOI provided above.

Line 380: This is an example of where a multiple regression model could be used.

Line 404: No mention was made in statistics of using nonlinear regressions