the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 15 May 2023
Data treatment and corrections for estimating H2O and CO2 isotope fluxes from high-frequency observations
Abstract. Current understanding of land-atmosphere exchange fluxes is limited by the fact that available observational techniques mainly quantify net fluxes, which are the sum of generally larger, bi-directional fluxes that partially cancel out. As a consequence, validation of gas exchange fluxes applied in models is challenging due to the lack of ecosystem-scale exchange flux measurements partitioned into soil, plant, and atmospheric components. One promising experimental method to partition measured turbulent fluxes uses the exchange-process-dependent isotopic fractionation of molecules like CO2 and H2O. When applying this method at a field scale, an isotope flux (δ-flux) needs to be measured. Here, we present and discuss observations made during the LIAISE 2021 field campaign using an Eddy Covariance (EC) system coupled to two laser spectrometers for high frequency measurement of the isotopic composition of H2O and CO2. This campaign took place in the summer of 2021 in the irrigated Ebro River basin near Mollerussa, Spain, embedded in a semi-arid region.
We present a systematic procedure to scrutinise and analyse the measured values of central δ-flux variable. Our experimental data indicated a larger relative signal loss in the δ-fluxes of H2O compared to the net ecosystem flux of H2O, while this was not true for CO2. Furthermore, we find that mole fractions and isotope ratios measured with the same instrument can be offset in time by more than a minute for the H2O isotopologues. We discuss how such artifacts can be detected and how they impact flux partitioning. We argue that these effects are likely due to condensation of water on a cellulose filter in our inlet system. Furthermore, we show that these artifacts can be resolved using physically sound corrections for inlet delays and high frequency loss. After such corrections and verification’s are made, isotopic ecosystem scale flux partitioning can be used reliably to validate conceptual land-atmosphere exchange models.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(6636 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(6636 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-785', Sonja Wahl, 06 Jul 2023
Dear authors,
Thank you for an interesting manuscript and a valuable dataset. I have a considerable amount of questions on details to the system set-up and the post-processing of the flux data. Please find my review in the attached pdf.
Best, Sonja Wahl
- AC1: 'Reply on RC1', Robbert Moonen, 18 Sep 2023
-
RC2: 'Comment on egusphere-2023-785', Anonymous Referee #2, 12 Jul 2023
Based on combined eddy-covariance and fast isotope measurements (with 2 different instruments) over flood irrigated alfalfa, the authors report case findings and methodological step innovations (especially a revision of an empirical, required response correction for the isotope timeseries).
General commentsThe manuscript well matches the scope of AMT and contributes to an important task, i.e. bringing forward combined isotope and EC measurements towards the point where they might eventually enable robust, continuously operational flux source partitioning. As far as the reviewer (who is more familiar with EC and micrometeorology than isotopes) can judge, the study design and analysis were thorough and state-of-the-art. The text and especially the figures are elaborated with care and mostly clear. A mixture of unfortunate small surprises (e.g. the suspected ambient air leak) and possibly oversights (if not overlooked, no actual partitioned fluxes are shown, which in turn might be due to the fact that no reference measurements of the isotopic composition of e.g. transpiration and evaporation water are mentioned) seems to have prevented the study from being a full success story about isotope-based EC flux partitioning and thus leave it somewhat behind the "cutting edge" of this area. It would be helpful for the readership to learn more clearly whether or not a complete source partitioning was the original aim and if so, what where the main obstacles. However, since mixed outcomes like this still appear to be the rule (the publication of which should help the community to learn about the prospects and pitfalls of the methodology) rather than the exception, and since the authors share interesting findings and innovations, it is definitely worthwhile publishing in AMT. That said, some issues mostly with the presentation require at least minor revisions. The style of documenting the methodology and offering explanations for findings has a slight tendency to be too "anecdotal", e.g. when mentioning a "reasonably short" inlet, claiming the insensitivity of results to the low-pass filter or using in the discussion's line of reasoning something that was looked at during data analysis but does not seem to be presented in the manuscript. More details on these examples and a few more are given below, but I would like to ask the authors to thoroughly screen the whole manuscript for such style, and keep it to a minimum - i.e. remove unnecessary undefended statements, defend (e.g. with data, calculations or appropriate citations) the necessary ones, and clarify the speculative status of the remaining ones.
Specific commentsL33-34: The short mentioning of global NEE partitioning together with the terms "the same principal" (princip*le*?) raises the expectation that the same methodology can be used for global and local isotope-based partitioning. However, on a local scale we are talking about partitioning measured fluxes (eddy covariance) with isofluxes, while on a global scale partitioning has to be inferred from concentrations, isoptope concentrations and additional assumptions. Depending on length issues and how important it is, I suggest to either change the wording such that no too tight identity between the two methods is expected, or alternatively explain the global method in a bit more detail, and then clearly mark where you start focusing on the local one relevant for your study.
L86-90: This part is obscure and at least near the end also not completely correct. Eq. (2) says nothing about gradients, it relates a covariance to a flux. A relation between any of both with the vertical gradient is not shown in the Eq. (and not in the Fig. alone either); it is subject to a constant micrometeorological variation of the proportionality "constant", and occasionally even counter-gradient fluxes do exist where the scale of transport-relevant eddies is larger than the vertical distance between local profile extrema (e.g. in or near canopies or with strong entrainment). Re-check which role (if any), knowledge of the gradients actually has in your methodology and where (my guess would be mostly if not everywhere) it is more appropriate to speak about fluxes or covariances. Even then gradients might help to "teach" your rationale in an idealized way (such as in Figure 1), but the reader should be able to easily separate these from cases where gradients are actually relevant to the derivation of your methodology (if any). Concerning delta-fluxes vs. isofluxes, try to find a more systematic way to relate both to each other and to the ultimately wanted partitioned fluxes. If the delta-Flux is the key part of the iso-flux as mentioned here, and the latter is needed for the partitioned flux, then which Equation term including Fdelta is the iso-flux? Or alternatively, if your chosen methodology circumvents any need to explicitly address the iso-flux, write this more clearly.
L115: As a first approximation literature typically assumes that ET it is always is limited by at least one of both, energy (in your case radiation) or water supply. The fact that both, radiation and soil water, were ample, does not automatically imply none of them was limiting. This is only the case if supplying even more could not increase ET more. Especially for radiation this does not necessarily need to be true in your case. It is acknowledged that third factors may limit ET instead, making it insensitive to changes in both water and radiation supply - such as e.g. turbulent removal or evaporated vapour, similar as discussed in Lobos-Roco et al. 2021 (Atm. Chem. Phys., https://doi.org/10.5194/acp-21-9125-2021). However, if you have such a non self-evident hypothesis it should be clarified in the manuscript, and whether it is supported by measurements or models for your site or merely an assumption. In most settings, more radiation tends to increase convection, such that ultimately it remains the limnitiung factor.
L133: The wording suggests that the authors are sure the pump flows add up arithmetically, ideally by having "closed the budget" with independent flow measurements of both components and the combination. If not sure, choose a wording that reverts to what you know for sure - e.g. how much flow the inlet scroll pump generates alone (with or without all the resistances of the complete system) and how much you measured in total for the combined system.
L141: "in principle" here means something like "in absence of the physical response issues we partly deal with in this study"? Note that it is good practice to still subject even open-path systems with 20 s-1 logging frequency to a correction for spectral loss - see also comment on L361. This may be negligible in the face of much larger loses of the isotope analyzers, and thus not necessary to apply in your study, partly also since the open-path system is an IRGASON with virtually no measurement path offset between w' and the concentrations (which however still suffer from some loss due to path length averaging). However, it should be avoided to accidently leave the impression that the mentioned logging frequencies alone should be sufficient, with the best available measurement system, to do without any spectral loss correction.
L186: Unclear statement (especially the "is described by") - do you want to say that assuming a 3 to 1 ratio approximately optimized the ability of a NL-tap-water to GL-icecore "virtual mixture" to mimick the atm measurement from Table 1?
L208: Much the same method is sometimes used to get rid of physical delays in EC data causing part of the spectral response issue, e.g. between sonic anemometer and gas analyzer (wind drift from one to the other, tube delays, other physical response time issues differeing between two instruments). What makes you think here that the result is "the most probably offset in the clocks"? If by design it includes the abovementioned effects, clarify whether this is a wanted (hopefully yes) or unwanted effect in this case.
L259: Can this be shown, even if only in the appendix?
L285, Fig. 8, Fig. 9 and others: The terms for the different instruments (e.g. OPGA, Aerodyne, Picarro, laser spectrometer) vary across the manuscript. Try to minimize variability to what is needed to stress the relevant physical properties of the instrument in the respective section / figure and, more important, to contrast natural pairs of terms (e.g. not the company name of one instrument with the measurement principle of the other instrument as in Fig. 8 and 9).
L323-325: "pose" is strange wording in this context - maybe "assume"? Which part of the reasoning is supported by the Gat and Yakir references how? In neither of them could find reference to the effect of altitude on drop formation, maybe because Gat is too long to be cited without a page reference. The bibliographic information on Gat at L496 is probably insufficient, check journal guidelines. Did you consider that the higher altitude at which droplets hit the surface in mountains does not necessarily imply a much higehr altitude of their formation?
Figure 9: The panel letters B to E referred to in the caption are missing in the figure, panel A is not referred to in the caption.
L357: The opening statement of the discussion is a bit bold given that the paper does not seem to try to actually show partitioned fluxes (e.g., transpiration and evaporation).
L361-364: For example, given this interesting question it would have been interesting to quantify the effect of the different correction methods on partitioned fluxes (e.g. on evaporation vs. transpiration). Open-path gas analyzer measurements are also still subject to spectral loss that requires correction (e.g. path averaging effect), is this applied or considered in any way (see comment on L141)?
L373: replace "made" by e.g. "analyzed", "showed", "plotted" or "considered" (the spectra are not made by the scientist but by nature and/or the instrument). Is this something only mentioned here or actually shown in the results, as the next sentence suggests? It contributes to the somnewhat "anecdotal" character of the manuscript.
L393: They do not affect the eddies, but the way in which they are seen in the data. Replacing "affect" by "concern" or "apply to" might also work.
L437: "The tool...": whole sentence strange, both language and (unclear) meaning
L426f: Try to make the outlook a bit less a collection of "claims" what your own work group intends in the future, and a bit more an open discussion of what the community including yourself can do as logical next steps based on your findings in this study (or their synthesis with those of others). Figure 10 and the discussion around it are interesting, however somewhat poorly embedded into what is already done (e.g. in the CO2 vs. H2O quadrant space). Though not sure, I wonder whether the outlook is actually the best (or even a proper) place to present such data new to the reader - what about either an own results or appendix subsection if there is enough material, or removing / keeping to a minimum for future studies?
Technical correctionsL10: Something missing in "of central delta-flux variable"
L43: group*s*
L226: exten*t*
L341 deform*s
L354: effect*
L391:-392; cause*, loss *of* covariance
L409: While *the*? principle
L436: analys*i*s
L455: presented *a*? methodological approach
Citation: https://doi.org/10.5194/egusphere-2023-785-RC2 - AC2: 'Reply on RC2', Robbert Moonen, 18 Sep 2023
-
AC3: 'Comment on egusphere-2023-785', Robbert Moonen, 09 Oct 2023
Best editor and reviewers,
We realize that our resubmitted manuscript contains a DOI link to our online accessible datasets which is not working.
https://doi.org/10.6084/m9.figshare.23828514.v2 is the correct link. Could this still be updated in the text?Thanks in advance and sorry for the inconvenience,
Robbert MoonenCitation: https://doi.org/10.5194/egusphere-2023-785-AC3
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-785', Sonja Wahl, 06 Jul 2023
Dear authors,
Thank you for an interesting manuscript and a valuable dataset. I have a considerable amount of questions on details to the system set-up and the post-processing of the flux data. Please find my review in the attached pdf.
Best, Sonja Wahl
- AC1: 'Reply on RC1', Robbert Moonen, 18 Sep 2023
-
RC2: 'Comment on egusphere-2023-785', Anonymous Referee #2, 12 Jul 2023
Based on combined eddy-covariance and fast isotope measurements (with 2 different instruments) over flood irrigated alfalfa, the authors report case findings and methodological step innovations (especially a revision of an empirical, required response correction for the isotope timeseries).
General commentsThe manuscript well matches the scope of AMT and contributes to an important task, i.e. bringing forward combined isotope and EC measurements towards the point where they might eventually enable robust, continuously operational flux source partitioning. As far as the reviewer (who is more familiar with EC and micrometeorology than isotopes) can judge, the study design and analysis were thorough and state-of-the-art. The text and especially the figures are elaborated with care and mostly clear. A mixture of unfortunate small surprises (e.g. the suspected ambient air leak) and possibly oversights (if not overlooked, no actual partitioned fluxes are shown, which in turn might be due to the fact that no reference measurements of the isotopic composition of e.g. transpiration and evaporation water are mentioned) seems to have prevented the study from being a full success story about isotope-based EC flux partitioning and thus leave it somewhat behind the "cutting edge" of this area. It would be helpful for the readership to learn more clearly whether or not a complete source partitioning was the original aim and if so, what where the main obstacles. However, since mixed outcomes like this still appear to be the rule (the publication of which should help the community to learn about the prospects and pitfalls of the methodology) rather than the exception, and since the authors share interesting findings and innovations, it is definitely worthwhile publishing in AMT. That said, some issues mostly with the presentation require at least minor revisions. The style of documenting the methodology and offering explanations for findings has a slight tendency to be too "anecdotal", e.g. when mentioning a "reasonably short" inlet, claiming the insensitivity of results to the low-pass filter or using in the discussion's line of reasoning something that was looked at during data analysis but does not seem to be presented in the manuscript. More details on these examples and a few more are given below, but I would like to ask the authors to thoroughly screen the whole manuscript for such style, and keep it to a minimum - i.e. remove unnecessary undefended statements, defend (e.g. with data, calculations or appropriate citations) the necessary ones, and clarify the speculative status of the remaining ones.
Specific commentsL33-34: The short mentioning of global NEE partitioning together with the terms "the same principal" (princip*le*?) raises the expectation that the same methodology can be used for global and local isotope-based partitioning. However, on a local scale we are talking about partitioning measured fluxes (eddy covariance) with isofluxes, while on a global scale partitioning has to be inferred from concentrations, isoptope concentrations and additional assumptions. Depending on length issues and how important it is, I suggest to either change the wording such that no too tight identity between the two methods is expected, or alternatively explain the global method in a bit more detail, and then clearly mark where you start focusing on the local one relevant for your study.
L86-90: This part is obscure and at least near the end also not completely correct. Eq. (2) says nothing about gradients, it relates a covariance to a flux. A relation between any of both with the vertical gradient is not shown in the Eq. (and not in the Fig. alone either); it is subject to a constant micrometeorological variation of the proportionality "constant", and occasionally even counter-gradient fluxes do exist where the scale of transport-relevant eddies is larger than the vertical distance between local profile extrema (e.g. in or near canopies or with strong entrainment). Re-check which role (if any), knowledge of the gradients actually has in your methodology and where (my guess would be mostly if not everywhere) it is more appropriate to speak about fluxes or covariances. Even then gradients might help to "teach" your rationale in an idealized way (such as in Figure 1), but the reader should be able to easily separate these from cases where gradients are actually relevant to the derivation of your methodology (if any). Concerning delta-fluxes vs. isofluxes, try to find a more systematic way to relate both to each other and to the ultimately wanted partitioned fluxes. If the delta-Flux is the key part of the iso-flux as mentioned here, and the latter is needed for the partitioned flux, then which Equation term including Fdelta is the iso-flux? Or alternatively, if your chosen methodology circumvents any need to explicitly address the iso-flux, write this more clearly.
L115: As a first approximation literature typically assumes that ET it is always is limited by at least one of both, energy (in your case radiation) or water supply. The fact that both, radiation and soil water, were ample, does not automatically imply none of them was limiting. This is only the case if supplying even more could not increase ET more. Especially for radiation this does not necessarily need to be true in your case. It is acknowledged that third factors may limit ET instead, making it insensitive to changes in both water and radiation supply - such as e.g. turbulent removal or evaporated vapour, similar as discussed in Lobos-Roco et al. 2021 (Atm. Chem. Phys., https://doi.org/10.5194/acp-21-9125-2021). However, if you have such a non self-evident hypothesis it should be clarified in the manuscript, and whether it is supported by measurements or models for your site or merely an assumption. In most settings, more radiation tends to increase convection, such that ultimately it remains the limnitiung factor.
L133: The wording suggests that the authors are sure the pump flows add up arithmetically, ideally by having "closed the budget" with independent flow measurements of both components and the combination. If not sure, choose a wording that reverts to what you know for sure - e.g. how much flow the inlet scroll pump generates alone (with or without all the resistances of the complete system) and how much you measured in total for the combined system.
L141: "in principle" here means something like "in absence of the physical response issues we partly deal with in this study"? Note that it is good practice to still subject even open-path systems with 20 s-1 logging frequency to a correction for spectral loss - see also comment on L361. This may be negligible in the face of much larger loses of the isotope analyzers, and thus not necessary to apply in your study, partly also since the open-path system is an IRGASON with virtually no measurement path offset between w' and the concentrations (which however still suffer from some loss due to path length averaging). However, it should be avoided to accidently leave the impression that the mentioned logging frequencies alone should be sufficient, with the best available measurement system, to do without any spectral loss correction.
L186: Unclear statement (especially the "is described by") - do you want to say that assuming a 3 to 1 ratio approximately optimized the ability of a NL-tap-water to GL-icecore "virtual mixture" to mimick the atm measurement from Table 1?
L208: Much the same method is sometimes used to get rid of physical delays in EC data causing part of the spectral response issue, e.g. between sonic anemometer and gas analyzer (wind drift from one to the other, tube delays, other physical response time issues differeing between two instruments). What makes you think here that the result is "the most probably offset in the clocks"? If by design it includes the abovementioned effects, clarify whether this is a wanted (hopefully yes) or unwanted effect in this case.
L259: Can this be shown, even if only in the appendix?
L285, Fig. 8, Fig. 9 and others: The terms for the different instruments (e.g. OPGA, Aerodyne, Picarro, laser spectrometer) vary across the manuscript. Try to minimize variability to what is needed to stress the relevant physical properties of the instrument in the respective section / figure and, more important, to contrast natural pairs of terms (e.g. not the company name of one instrument with the measurement principle of the other instrument as in Fig. 8 and 9).
L323-325: "pose" is strange wording in this context - maybe "assume"? Which part of the reasoning is supported by the Gat and Yakir references how? In neither of them could find reference to the effect of altitude on drop formation, maybe because Gat is too long to be cited without a page reference. The bibliographic information on Gat at L496 is probably insufficient, check journal guidelines. Did you consider that the higher altitude at which droplets hit the surface in mountains does not necessarily imply a much higehr altitude of their formation?
Figure 9: The panel letters B to E referred to in the caption are missing in the figure, panel A is not referred to in the caption.
L357: The opening statement of the discussion is a bit bold given that the paper does not seem to try to actually show partitioned fluxes (e.g., transpiration and evaporation).
L361-364: For example, given this interesting question it would have been interesting to quantify the effect of the different correction methods on partitioned fluxes (e.g. on evaporation vs. transpiration). Open-path gas analyzer measurements are also still subject to spectral loss that requires correction (e.g. path averaging effect), is this applied or considered in any way (see comment on L141)?
L373: replace "made" by e.g. "analyzed", "showed", "plotted" or "considered" (the spectra are not made by the scientist but by nature and/or the instrument). Is this something only mentioned here or actually shown in the results, as the next sentence suggests? It contributes to the somnewhat "anecdotal" character of the manuscript.
L393: They do not affect the eddies, but the way in which they are seen in the data. Replacing "affect" by "concern" or "apply to" might also work.
L437: "The tool...": whole sentence strange, both language and (unclear) meaning
L426f: Try to make the outlook a bit less a collection of "claims" what your own work group intends in the future, and a bit more an open discussion of what the community including yourself can do as logical next steps based on your findings in this study (or their synthesis with those of others). Figure 10 and the discussion around it are interesting, however somewhat poorly embedded into what is already done (e.g. in the CO2 vs. H2O quadrant space). Though not sure, I wonder whether the outlook is actually the best (or even a proper) place to present such data new to the reader - what about either an own results or appendix subsection if there is enough material, or removing / keeping to a minimum for future studies?
Technical correctionsL10: Something missing in "of central delta-flux variable"
L43: group*s*
L226: exten*t*
L341 deform*s
L354: effect*
L391:-392; cause*, loss *of* covariance
L409: While *the*? principle
L436: analys*i*s
L455: presented *a*? methodological approach
Citation: https://doi.org/10.5194/egusphere-2023-785-RC2 - AC2: 'Reply on RC2', Robbert Moonen, 18 Sep 2023
-
AC3: 'Comment on egusphere-2023-785', Robbert Moonen, 09 Oct 2023
Best editor and reviewers,
We realize that our resubmitted manuscript contains a DOI link to our online accessible datasets which is not working.
https://doi.org/10.6084/m9.figshare.23828514.v2 is the correct link. Could this still be updated in the text?Thanks in advance and sorry for the inconvenience,
Robbert MoonenCitation: https://doi.org/10.5194/egusphere-2023-785-AC3
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 256 | 98 | 25 | 379 | 15 | 13 |
- HTML: 256
- PDF: 98
- XML: 25
- Total: 379
- BibTeX: 15
- EndNote: 13
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Robbert Petrus Johannes Moonen
Getachew Agmuas Adnew
Oscar Karel Hartogensis
Jordi Vilà-Guerau de Arellano
David Joan Bonell Fontas
Thomas Röckmann
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(6636 KB) - Metadata XML