Decision on MS BG 2017-56

Dear authors,

I have now read your answers to the reviewer’s comments and projected changes to the manuscript. Because I found reviewer’s comments were at some places superficial, I made my own detailed reading of your submitted paper. Although it was clear to me that the presented dataset is of high quality for publication, my major concern comes from the lack of well-defined objectives of the paper and the justification of the choices made in the way data are analysed and presented. What are the important new insights provided by the study? How can this paper help future CO2 and CH4 fluxes measurements in lakes? These questions must be answered more clearly in your revised MS. It was not evident to me if your paper was focussed on the comparison between EC, FC and k methods, on the validity of different parameterization of k, on the diurnal variations of fluxes at the study site, or on the difference between stratified and mixed periods at the study site, etc…

To that respect, I recommend you consider very seriously the general comment by Ref.#1: “some aspects of the discussion elements should be reworded to make the main conclusions more evident. Some of the conclusions, for example on the difference between day and night time, do not seem so evident based on the figure analysis. Those figures should also be improved to ease comparisons between fluxes and controlling parameters on one side, and between approaches on the other side.”

I also fully agree with review#1’s comment “If you have floating chamber measurements and mixing ratio of CO2 and CH4 in the water, you should try to calculate a site specific k value and compare it to the one calculated from Cole and Caraco”
I would add that k could be calculated from EC and FC fluxes and compared to those calculated with the different BLM. If the paper deals with performance of methods (maybe not), calculations of k can be done separately for CO2 and CH4, normalised with the Schmidt Number and compared one with the other. This would allow a comparison of performance of methods, even when fluxes are extremely low for CH4 but not for CO2 (due to larger delta CO2 between air and water). In your respond to this comment, you mention a second paper in preparation. Your dataset is indeed strong enough to produced two papers, but splitting the data in two papers must be done according to well-defined specific objectives.

Five Figures (6-10) report “median” diurnal variations of fluxes and driving parameters. If I understood properly, these are composite data averaging values at the same hour but different days, so these data include natural day-to-day variations. I was not convinced that using such composite median values was the best way of comparing the efficiency of the different methods, which could be done using individual synchronous values.

To summarise, I am happy to encourage you to proceed with the revision of your manuscript, but I would like to stress that this should be done after re-think profoundly the objective of your study and the way the data are analysed. Needs some work. The paper must be re-focused. Separating the results section from the discussion section will probably help your writing.
One detail: P4L27-Final gas concentrations in the water were calculated using the Henry’s Law
Please mention what solubility coefficients you used for CO2 and CH4

Looking forward to reading a revised version of you MS.

With best regards
Gwenaël Abril, BG Associate Editor

Second review of BG 2017 58

Dear authors

Because I wanted to move forward the edition of your paper, I didn’t send it back to reviewers and I made my own detailed reading of your revised MS. My conclusion is that the presented dataset is of excellent quality and your conclusions are highly relevant for the community. Your paper must be published soon in BG. However, this large amount of data must be better organized, presented and synthesized in order to be published as a paper. I still find your revised MS confusing and difficult to read and unfortunately, I cannot recommend publication in its present form. When compared to the initial submission, little changes were made in the organization of the MS and some of my previous comments are still valid with this revised version. The result and discussion section is the most critical part of your MS, being often confusing and difficult to follow; fortunately the abstract and the conclusion are excellent and, together with figures Figure 2-5, they convince the reader of the quality of the presented work. However, throughout your “result and discussion section” you jump from CH4 fluxes to CO2 fluxes, from one method to another, from the stratified to mixed period, from discussion of methods to discussion of processes, etc… As you show and discuss your data at the same time, the reader often gets lost. I respect your choice to maintain results and discussion together; however, the organization of the text does not always appear logical (see below). In addition some figures are also confusing, particularly the five figures 6-10 showing diurnal variations of different averaged parameters, in many cases using different axes which complicates even more the reading and the comparison of one panel with another; The information contained in these 5 repetitive figures must be summarized in one or two figures showing the most important findings and a couple of tables with detailed statistical analysis (are fluxes for stratified/mixed, daytime/nightime and different methods, significantly different or not ?). Alternatively, these 5 figures could appear as Suppl. Material. Any change that will make the result and discussion shorter, more focussed and easier to read will be welcome.

Content of the result and discussion section is as follows:

3.1 Environmental conditions
3.1.1 Water column temperature and gas profiles
Here CH4 and CO2 are mixed
Note that there is no sub-section 3.1.2

3.2 Comparison between boundary layer model and eddy covariance flux estimates
3.2.1 CH4 fluxes
3.2.2 CO2 fluxes
3.3 Diurnal variation of estimated fluxes
3.3.1 Stratified period
3.3.2 Mixing period
3.4 Comparison between floating chambers and eddy covariance fluxes
3.4.1 CH4 fluxes
3.4.2 CO2 fluxes

What is the motivation for comparing BLM and EC in 3.2 and FC and EC in 3.4, with diurnal variation in between? Why not comparing BLM with FC?
When diurnal variation is not significant it is probably not necessary to show the curves.

P8 section 3.11 and throughout the whole MS: Concentration units are sometimes mmol m-3 and sometimes nmol m-3. I find strange µmol m-3 never appear. Please check. I suggest choosing one single concentration unit for each gas.

P8
L20 : “equilibration time of 40 min should be enough” : any objective evidence for that?
L22-24: discuss the effect of concentration on calculated BLM fluxes
L26-27 “rapid increase in the surface water concentration according to manual sample” rephrase
L30 “agree better” : how much ?
P9
L14 “higher than” refer to a table with statistical analysis that compare all types of fluxes
P10 L14-15 “tested with mann-whitney U test” refer to a table with statistical analysis that compare all types of fluxes
L18 (about 60%) be precise
L25 CH4 appears in a section dedicated to CO2
L28-35: These important discussion statements appear very diluted within the more systematic description of results appearing above.
P11
L4-6. Is comparison between methods the only interest for describing diurnal variation?
L14 “negligible” : provide a value
L16 “the convective term in kHE increases toward night-time causing higher total kHE” confusing, rephrase
L22-23: “highest fluxes at noon when also friction velocity gains its maximum value” also when CO2 concentration was lower? The convective term is on a different scale in fig7, how much does it contribute at max and on average ?
L27 “lower u*w” lower than what ?
P12
“because the afternoon flux peak is also seen in the BLM by kCC, we can deduce that it is due to higher wind speed and enhanced shear during the afternoon as well as CH4 concentration difference”: you have CH4 concentration and wind speed data, no need to deduce from the value of kCC
L7-8: because you are discussing methods and processes at the same time, the text becomes hard to follow. Also cite Dumestre et al. 1999 Influence of Light Intensity on Methanotrophic Bacterial Activity in Petit Saut Reservoir, French Guiana APPLIED AND ENVIRONMENTAL MICROBIOLOGY,0099-2240/99/$04.0010Feb. 1999, p. 534–539
“the larger cc difference toward the afternoon may be caused by higher oxidation rate in the dark … during night” rephrase
L14-15 “all models give similar diurnal patterns… only magnitude are different”. Ok, but this is due to the predominant effect of changes in CH4 concentrations
L17-21: discussion in the light of literature could be strengthened if it did not appear only as separated statements at the end of each paragraph.
L29 “which is then visible” rephrase being more precise
L30-31, indeed, but kCC agrees with EC during part of the day. How relevant are these comparisons based on averages of measurement during 5 following days (figures 6,8 & 10): what about variation from one day to another?
P13
L 3-5 “using selectively only daytime gas concentration… global budget makes a biased assumption” AND “the EC… no clear diurnal variation during this period either”. This looks like a contradiction
L14 “and the coefficient of variation was…” provide average +/- coefficient of variation and avoid such phrasing
L19 “partly this difference is of course due to FC fluxes averaged over the different measurements spots… “ rephrase
L24 “low fluxes are difficult to detect… close to the detection limit of the gas analyser used in the EC measurements” The detection limits of a gas analyser concerns concentration and not fluxes. Gas analysers are able to measure standard atmospheric concentration. Here you may reach the limit for EC fluxes calculation, depending on various classical criteria of EC data processing (spectral analysis, stability, etc...) Please provide this information.
L25 “could have probably produced a better comparison” unclear statement
L28 “statistically different” provide a table with complete statistical analysis. What means “to detect reliability” ?
L19 “larger source area” rephrase
P14
L4 “differed from daytime EC fluxes” provide a table with detailed statistical analysis
L8-10 referring to literature elsewhere only at the end of paragraphs makes the discussion superficial
L12: provide a table with detailed statistical analysis
L17 you are discussing CH4 fluxes in a section dealing with CO2 fluxes
L21 “this is due to limitations in the EC method” do you mean “this is one of the limitations…”?

To summarize, I find your work and conclusions of excellent scientific quality, but your MS needs substantial improvement so readers can access more easily to your dataset and conclusions. I will be happy to receive a new MS with substantial revision of the results and discussion section, figures and tables.
Best Regards
Gwenaël Abril

Dear authors, I am pleased to accept your paper for publication in Biogeosciences. I indicate here some technical points you should consider before sending the definitive version of your MS for publication

I appreciated the effort with the figures and Tables, data presentation has greatly improved. The two diurnal cycles you have selected (from the 4 in the previous version) as Sup. Mat in this revised version provide significant information and are well cited in the text. At your convenience, you may insert Figs A1 and A2 so they can be published in the MS, which is now limited to 6 figures and there might be some room for 8.

P9-12 Numbering of the discussion sections (3.3 and 3.3.1 – 3.4 and 3.4.1) are a bit strange. Indeed, to be useful, subtitle 3.3.1 would require at least a sub-titles 3.3.2, eventually a 3.3.3…
A simplest option would be to remove 3.3.1 and 3.4.1 if you agree to include spatial variations in “CO2 flux comparison”.
Insert “a” and “b” in the two panels of figure 6 (and 4)