Eddy-covariance carbon fluxes of a heterogeneous forest: one tower - two heights

Krasnova, Alisa; Krasnov, Dmitrii; Cordey, Hans Peter Ernst; Noe, Steffen M.

doi:https://doi.org/10.5194/egusphere-2022-384

Preprints

https://doi.org/10.5194/egusphere-2022-384

Preprints

09 Jun 2022

| 09 Jun 2022

Eddy-covariance carbon fluxes of a heterogeneous forest: one tower - two heights

Alisa Krasnova, Dmitrii Krasnov, Hans Peter Ernst Cordey, and Steffen M. Noe

Abstract. Eddy-covariance (EC) is a widely used method for measuring ecosystem-scale fluxes of various gases. The sensor placement height is typically constrained by the canopy height and area of interest size. We studied the carbon dioxide fluxes over a hemiboreal mixed forest with two EC measurement systems located at 30 m and 70 m. The lower system NEE (NEE30) values were more positive (smaller sink or higher source) than the NEE of the higher one (NEE70), but this difference was prevalent in low light conditions and in May–November of all studied years. The nighttime and early morning difference (ΔNEE) increased with wind speed until ~2 m s^-1 and friction velocity until ~0.35 m s^-1 and linearly decreased after. ΔNEE was irregularly distributed over the wind direction sectors with high values overlapping the directions of South-East and South-West guy wire tunnels. Moreover, the shape of the NEE30 seasonal cycle was closer to that of a clear-cut area, and the difference between the systems increased with air temperature. The forest under study varied between a weak net sink and a strong net source on the annual scale. Directional heterogeneity correction shifted the annual NEE towards more negative values, but neither removed the difference between the systems nor changed the shape of the seasonal cycle. More studies are needed to assess the impact of clearcutting on the carbon accumulation under the measurement point.

Received: 23 May 2022 – Discussion started: 09 Jun 2022

Download & links

Alisa Krasnova, Dmitrii Krasnov, Hans Peter Ernst Cordey, and Steffen M. Noe

Status: closed

RC1:
'Comment on egusphere-2022-384', Anonymous Referee #1, 27 Jun 2022

General comments

The question of the impact of source area heterogeneity on carbon (and other species) fluxes is highly relevant and interesting. Flux sites are often chosen where the heterogeneity is as small as possible but in reality, there are always some heterogeneity to consider. The title is however a bit misleading – it is not only heterogeneity of a forest but of a landscape. The first thought when reading such a title is that this must deal a lot with land cover and footprint analyses. But this is not the case. It is mainly a description of differences in fluxes in different wind sectors without any attempt to quantify differences in land cover and footprints in relation to these land cover variations and how this might impact the fluxes.

It is well known that the footprint area varies a lot depending on day/night unstable/stable conditions but this is hardly mentioned or only indirectly. Some of the figures, for instance Fig. 7e&f, where the difference in NEE between the two heights are plotted against u* and/or u for two different radiation levels gives a strong indication that stability has a large impact on the difference.

Another issue which might have impact on the fluxes are vertical and horizontal advection. It is not clear from the methods if there are concentration measurements at different heights in the tower but there are some mentioning of ‘plumbing’ which indicate that there might be such data available. If it is, then also total advection could be estimated. See for instance Yi et al Influence of advection on measurements of the net ecosystem-atmosphere exchange of CO2 from a very tall tower JGR Vol. 105, No. D8, 9991-9999, 2000.

The conclusion that it is the carbon enriched air from the clear cutting along the guy wires that are causing the spatial pattern of CO2 fluxes that are observed for the 30 m system is speculative without firm evidence. In the conclusion they also state that their hypothesis is confirmed and with the hypothesis that “the EC system located closer to the canopy will detect local features while the higher positioned system will detect naturally integrated CO2 flux”, this is obvious and no need to be ‘confirmed’.

Specific comments

Line 67-68: Describe the type of tower that is used. Please specify the width of the clearing along the guy wires. It is hardly visible in Fig. 2. You don’t mention if you have removed data from situations when the wind come through the tower. It would be good if you could indicate boom direction in some of your figures as well.

L 75: You present measurements over four years (which is good). Please state how often the gas analyzer was calibrated.

L 81: Here you mention that ‘soil efflux’ is measured with a transparent chamber. Later you refer to these measurements as ‘soil respiration’. If you are using a transparent chamber then you are measuring net exchange rather than respiration. Please clarify.

L 87: No measurement of net radiation? You don’t say anything about energy balance closure which is a common kind of quality measure for flux measurements.

L 106: Net ecosystem exchange NEE should include the storage term as well. You should make this clear if it was not and then name the flux otherwise.

L 155: Why not write out V/A instead of ‘h’ in eq. 4 to avoid the reader from reacting.

L 195 & Fig. 3. I don’t see that the wind profile is used at all anywhere in the paper. So you can remove it.

L294-296: Here it become important to know if you really measured ‘soil respiration’ or net flux from the forest floor.

L 375-376: There are papers on multi-level EC measurements from high towers. For instance the one mentioned above by Yi et al. and Davis et al. Global Change Biology (2003) 9, 1278–1293, Berger et al. Journal of Atmospheric and Oceanic Technology, Vol. 18, 529-542, 2001.

Fig. 2. It would be much more informative if the footprint climatology was divided in night & day and if more isolines were presented (e.g. 20,40, 60, 80%). And a map of land uses would also be more informative than an aerial photograph (which also can be shown for itself).

Fig. 11-12: here it would be really interesting with a proper footprint analyses.

Citation: https://doi.org/10.5194/egusphere-2022-384-RC1
- AC1: 'Reply on RC1', Alisa Krasnova, 12 Aug 2022
  
  We would like to thank the Anonymous Referee #1 for their time. Our answers are marked in bold
  General comments
  The question of the impact of source area heterogeneity on carbon (and other species) fluxes is highly relevant and interesting. Flux sites are often chosen where the heterogeneity is as small as possible but in reality, there are always some heterogeneity to consider. The title is however a bit misleading – it is not only heterogeneity of a forest but of a landscape. The first thought when reading such a title is that this must deal a lot with land cover and footprint analyses. But this is not the case. It is mainly a description of differences in fluxes in different wind sectors without any attempt to quantify differences in land cover and footprints in relation to these land cover variations and how this might impact the fluxes.
  Thank you for this comment! We will add footprint and land cover analysis to the revised manuscript. We also will consider changing the title.
  It is well known that the footprint area varies a lot depending on day/night unstable/stable conditions but this is hardly mentioned or only indirectly. Some of the figures, for instance Fig. 7e&f, where the difference in NEE between the two heights are plotted against u* and/or u for two different radiation levels gives a strong indication that stability has a large impact on the difference.
  Yes, you are correct. We will assess the impact of stability on the difference between the heights
  Another issue which might have impact on the fluxes are vertical and horizontal advection. It is not clear from the methods if there are concentration measurements at different heights in the tower but there are some mentioning of ‘plumbing’ which indicate that there might be such data available. If it is, then also total advection could be estimated. See for instance Yi et al Influence of advection on measurements of the net ecosystem-atmosphere exchange of CO2 from a very tall tower JGR Vol. 105, No. D8, 9991-9999, 2000.
  Thank you for the recommendation. We will add total advection estimation (we have concentration profile measurements at 30, 50 and 70 m)
  The conclusion that it is the carbon enriched air from the clear cutting along the guy wires that are causing the spatial pattern of CO2 fluxes that are observed for the 30 m system is speculative without firm evidence. In the conclusion they also state that their hypothesis is confirmed and with the hypothesis that “the EC system located closer to the canopy will detect local features while the higher positioned system will detect naturally integrated CO2 flux”, this is obvious and no need to be ‘confirmed’.
  Thank you, we will revise the conclusions
  Specific comments
  Line 67-68: Describe the type of tower that is used. Please specify the width of the clearing along the guy wires. It is hardly visible in Fig. 2. You don’t mention if you have removed data from situations when the wind come through the tower. It would be good if you could indicate boom direction in some of your figures as well.
  The information will be added to the revised manuscript
  L 75: You present measurements over four years (which is good). Please state how often the gas analyzer was calibrated.
  The information will be added to the revised manuscript
  L 81: Here you mention that ‘soil efflux’ is measured with a transparent chamber. Later you refer to these measurements as ‘soil respiration’. If you are using a transparent chamber then you are measuring net exchange rather than respiration. Please clarify.
  Yes, our soil chambers are transparent, meaning that the obtained fluxes represent soil floor respiration. After careful consideration we decided to omit soil fluxes from this study, due to their minor input to the final conclusions.
  L 87: No measurement of net radiation? You don’t say anything about energy balance closure which is a common kind of quality measure for flux measurements.
  The station is lacking soil heat flux measurements, so the energy balance closure estimation is not possible.
  L 106: Net ecosystem exchange NEE should include the storage term as well. You should make this clear if it was not and then name the flux otherwise.
  The information will be added to the revised manuscript
  L 155: Why not write out V/A instead of ‘h’ in eq. 4 to avoid the reader from reacting.
  Thank you for the suggestion. We decided to omit soil fluxes from this study, due to their minor input to the final conclusions.
  L 195 & Fig. 3. I don’t see that the wind profile is used at all anywhere in the paper. So you can remove it.
  Thank you, the figure will be removed
  L294-296: Here it become important to know if you really measured ‘soil respiration’ or net flux from the forest floor.
  Thank you for the suggestion. We decided to omit soil fluxes from this study, due to their minor input to the final conclusions.
  L 375-376: There are papers on multi-level EC measurements from high towers. For instance the one mentioned above by Yi et al. and Davis et al. Global Change Biology (2003) 9, 1278–1293, Berger et al. Journal of Atmospheric and Oceanic Technology, Vol. 18, 529-542, 2001.
  Thank you for the recommendation!
  Fig. 2. It would be much more informative if the footprint climatology was divided in night & day and if more isolines were presented (e.g. 20,40, 60, 80%). And a map of land uses would also be more informative than an aerial photograph (which also can be shown for itself).
  The map will be remade in the revised version of the manuscript
  Fig. 11-12: here it would be really interesting with a proper footprint analyses.
  The footprint analysis will be added to the revised manuscript
  
  Citation: https://doi.org/10.5194/egusphere-2022-384-AC1
RC2:
'Comment on egusphere-2022-384', Ankur Desai, 27 Jun 2022

Krasnova and colleagues evaluate sources of differences in eddy covariance fluxes made at two heights, 30 and 70 m on a tall tower at the SMEAR Estonia tower site. They conclude that the lower height is more affected by nearby clearings, leading to a smaller sink or larger sources, especially in calm conditions. Further, use of a wind directional rectification approach did not significantly alter this effect but allowed for estimation of a carbon budget in line with expectations. Overall, the contribution is useful for understanding the importance of evaluating sources of footprint bias and approaches to account for those in making reliable, defensible flux measurements over heterogenous ecosystems.

While the concept is useful, the manuscript lacks key consideration of drivers of flux profile in the atmospheric surface and boundary-layer. An eddy flux measured at a specific height (Fc) is not necessarily equivalent to the surface flux (NEE). Rather it is the flux at that height, and most flux profiles have linear slopes in the boundary-layer beyond the lower surface layer, of which 70 m would often not be part. So it should be no surprise that the two fluxes differ by some magnitude. Without accounting for flux divergence, advection, and storage flux, it is not clear that the two heights can be compared. My recommendation is thus the manuscript be rejected now, but a resubmission that addresses this oversight would be a positive contribution to the literature.

1.) In particular, the authors appear to have failed to account for the known effect of storage flux, the accumulation of flux under sensor height. Storage flux underestimation would be consistent with the delta_NEE (30-70) term being positive under a strong surface source, and during calm wind, low u* conditions at low light or night.

This is something I showed at our own 396 m tall tower site with fluxes measured at 30, 122, and 396 m for methane and CO2 fluxes in Desai et al., 2015. I encourage the authors to take a look and see if a similar analysis would be possible here using the CO2 concentration profile, even just from the two heights. Similarly, this particular site (Ameriflux/Fluxnet site US-PFa) has been operating since 1996 and there have been several publications looking at flux profiles and differences with height, including Davis et al., 2003, Berger et al., 2001, and Yi et al., 2000. It's too bad these were missed. Note, however, I do not require the authors to cite any work I mention - but hopefully these papers give you some guidance on how to approach the comparison here from a similar type of set up.

2.) I don't doubt that footprint differences are also a driving factor, but without an analysis of the flux profile and including storage flux (a known issues for tall towers back to foundational work on Massman and Lee, 2002), it is difficult to truly test your hypothesis. Beyond the storage flux issue, there is also the relatively qualitative way the authors try to evaluate the role of clearings and land surface with wind direction at the two heights. Since the authors have already run the Kljun 2D footprint model to construct the footprint climatology, I recommend developing a simple land cover map for both heights and evaluating the relative contributions to the footprint climatology. Also, given that the delta_NEE signal does have a wind directional signal, it would seem wise to try to look at the specific footprints during periods of max difference and identify whether the land cover fractions differ significantly (and whether the clearing is a major influence on the 30 m level).

I wonder if another way to test for this is to look at the same differences for CO2 but for H2O or heat flux? One might expect that lower GPP element from the clearing is driving the flux difference, one would also see differences in higher heat fluxes (more surface heating) and lower water fluxes (less transpiration). This is not a required correction, but could be a fun way to further support the hypothesis.

3.) The use of the chamber fluxes is relatively minimal. Personally, I find it provides compelling evidence that EC70 is more representative of the landscape (assuming the chambers are representative of the landscape), as both the magnitude and pattern more closely resemble EC70. I would suspect, given the forest cover here, that Rs would be more on the higher end of Rs/ER ratio report (0.3-0.8). Among other things the cited paper by Barba et al demonstrates that at many sites RSoil often exceeds REco, or is very similar in magnitude (look at Table 3 and the conclusion within it). It's not clear to me where the 0.3 to 0.8 range is actually mentioned in the citations. Davidson and Janssens makes no mention of ratios as far as I can tell. Either more work needs to be done to fully incorporate analysis of chamber fluxes, or it should be removed.

4.) The wind direction normalization from Hadden and Grelle (2017, also cited as 2016, is that a different paper?) seems like a good first test for estimating annual budgets, but there have been some more advanced approaches in recent years. For this manuscript, since the goal is just to demonstrate how consistent budgets are after normalization, probably not a major issues. But take a look at Griebel et al., 2020, Wang et al., 2006, Metzger, 2018.

Citations:

Berger, B. W., Davis, K. J., Yi, C., Bakwin, P. S., & Zhao, C. L. (2001). Long-Term Carbon Dioxide Fluxes from a Very Tall Tower in a Northern Forest: Flux Measurement Methodology, Journal of Atmospheric and Oceanic Technology, 18(4), 529-542. https://doi.org/10.1175/1520-0426(2001)018<0529:LTCDFF>2.0.CO;2

Davis, K.J., Bakwin, P.S., Yi, C., Berger, B.W., Zhao, C., Teclaw, R.M. and Isebrands, J.G. (2003), The annual cycles of CO2 and H2O exchange over a northern mixed forest as observed from a very tall tower. Global Change Biology, 9: 1278-1293. https://doi.org/10.1046/j.1365-2486.2003.00672.x

Desai, A.R., Xu, K., Tian H., Weishampel, P., Thom, J., Baumann, D., Andrews, A.E., Cook, B.D., King, J.Y., and Kolka, R., 2015. Landscape-level terrestrial methane flux observed from a very tall tower. Agricultural and Forest Meteorology, 201, 61-75, doi:10.1016/j.agrformet.2014.10.017.

Griebel, A., Metzen, D., Pendall, E., Burba, G., & Metzger, S. (2020). Generating spatially robust carbon budgets from flux tower observations. Geophysical Research Letters, 47, e2019GL085942. https://doi.org/10.1029/2019GL085942

Massman, W. J., & Lee, X. (2002). Eddy covariance flux corrections and uncertainties in long-term studies of carbon and energy exchanges. Agricultural and Forest Meteorology, 113(1), 121-144. doi:https://doi.org/10.1016/S0168-1923(02)00105-3

Metzger, 2018. Surface-atmosphere exchange in a box: Making the control volume a suitable representation for in-situ observations. AGricultural and Forest Meteorology, 255, 68-80, https://doi.org/10.1016/j.agrformet.2017.08.037

Wang, W., Davis, K. J., Cook, B. D., Butler, M. P., and Ricciuto, D. M. (2006), Decomposing CO2 fluxes measured over a mixed ecosystem at a tall tower and extending to a region: A case study, J. Geophys. Res., 111, G02005, doi:10.1029/2005JG000093.

Yi, C., Davis, K. J., Bakwin, P. S., Berger, B. W., and Marr, L. C. (2000), Influence of advection on measurements of the net ecosystem-atmosphere exchange of CO2 from a very tall tower, J. Geophys. Res., 105( D8), 9991– 9999, doi:10.1029/2000JD900080.

Sincerely,

Ankur Desai, Professor, UW-Madison

Citation: https://doi.org/10.5194/egusphere-2022-384-RC2
- AC2: 'Reply on RC2', Alisa Krasnova, 12 Aug 2022
  
  On behalf of all co-authors, I would like to thank you for the review and all the recommendations given. Please find our answers marked in bold.
  Krasnova and colleagues evaluate sources of differences in eddy covariance fluxes made at two heights, 30 and 70 m on a tall tower at the SMEAR Estonia tower site. They conclude that the lower height is more affected by nearby clearings, leading to a smaller sink or larger sources, especially in calm conditions. Further, use of a wind directional rectification approach did not significantly alter this effect but allowed for estimation of a carbon budget in line with expectations. Overall, the contribution is useful for understanding the importance of evaluating sources of footprint bias and approaches to account for those in making reliable, defensible flux measurements over heterogenous ecosystems.
  While the concept is useful, the manuscript lacks key consideration of drivers of flux profile in the atmospheric surface and boundary-layer. An eddy flux measured at a specific height (Fc) is not necessarily equivalent to the surface flux (NEE). Rather it is the flux at that height, and most flux profiles have linear slopes in the boundary-layer beyond the lower surface layer, of which 70 m would often not be part. So it should be no surprise that the two fluxes differ by some magnitude. Without accounting for flux divergence, advection, and storage flux, it is not clear that the two heights can be compared. My recommendation is thus the manuscript be rejected now, but a resubmission that addresses this oversight would be a positive contribution to the literature.
  Thank you for the valuable recommendation! We will revise the manuscript according to your suggestions.
  1.) In particular, the authors appear to have failed to account for the known effect of storage flux, the accumulation of flux under sensor height. Storage flux underestimation would be consistent with the delta_NEE (30-70) term being positive under a strong surface source, and during calm wind, low u* conditions at low light or night.
  In the absence of storage measurement system, the storage term was calculated with the simplified method using only concentrations at the respective heights. Unfortunately, this information was missing from the methods section. We will add it to the revised manuscript.
  This is something I showed at our own 396 m tall tower site with fluxes measured at 30, 122, and 396 m for methane and CO2 fluxes in Desai et al., 2015. I encourage the authors to take a look and see if a similar analysis would be possible here using the CO2 concentration profile, even just from the two heights. Similarly, this particular site (Ameriflux/Fluxnet site US-PFa) has been operating since 1996 and there have been several publications looking at flux profiles and differences with height, including Davis et al., 2003, Berger et al., 2001, and Yi et al., 2000. It's too bad these were missed. Note, however, I do not require the authors to cite any work I mention - but hopefully these papers give you some guidance on how to approach the comparison here from a similar type of set up.
  We appreciate your recommendations and will update our analysis for the revised version of the manuscript.
  2.) I don't doubt that footprint differences are also a driving factor, but without an analysis of the flux profile and including storage flux (a known issues for tall towers back to foundational work on Massman and Lee, 2002), it is difficult to truly test your hypothesis. Beyond the storage flux issue, there is also the relatively qualitative way the authors try to evaluate the role of clearings and land surface with wind direction at the two heights. Since the authors have already run the Kljun 2D footprint model to construct the footprint climatology, I recommend developing a simple land cover map for both heights and evaluating the relative contributions to the footprint climatology. Also, given that the delta_NEE signal does have a wind directional signal, it would seem wise to try to look at the specific footprints during periods of max difference and identify whether the land cover fractions differ significantly (and whether the clearing is a major influence on the 30 m level).
  Thank you for this suggestion! The footprint analysis will be added to the revised manuscript.
  I wonder if another way to test for this is to look at the same differences for CO2 but for H2O or heat flux? One might expect that lower GPP element from the clearing is driving the flux difference, one would also see differences in higher heat fluxes (more surface heating) and lower water fluxes (less transpiration). This is not a required correction, but could be a fun way to further support the hypothesis.
  Thank you, we will definitely try this!
  3.) The use of the chamber fluxes is relatively minimal. Personally, I find it provides compelling evidence that EC70 is more representative of the landscape (assuming the chambers are representative of the landscape), as both the magnitude and pattern more closely resemble EC70. I would suspect, given the forest cover here, that Rs would be more on the higher end of Rs/ER ratio report (0.3-0.8). Among other things the cited paper by Barba et al demonstrates that at many sites RSoil often exceeds REco, or is very similar in magnitude (look at Table 3 and the conclusion within it). It's not clear to me where the 0.3 to 0.8 range is actually mentioned in the citations. Davidson and Janssens makes no mention of ratios as far as I can tell. Either more work needs to be done to fully incorporate analysis of chamber fluxes, or it should be removed.
  After careful consideration we decided to omit soil fluxes from this study due to their minor input to the final conclusions.
  4.) The wind direction normalization from Hadden and Grelle (2017, also cited as 2016, is that a different paper?) seems like a good first test for estimating annual budgets, but there have been some more advanced approaches in recent years. For this manuscript, since the goal is just to demonstrate how consistent budgets are after normalization, probably not a major issues. But take a look at Griebel et al., 2020, Wang et al., 2006, Metzger, 2018.
  The simple approach by Hadden and Grelle (yes, they have 2 different papers, one in 2016 and another one is 2017) was applied to check if the observed differences were caused by the irregular distribution of wind direction. We will also look at the suggested publications.
  
  Citation: https://doi.org/10.5194/egusphere-2022-384-AC2
CC1: 'Comment on egusphere-2022-384', Matthew Hayek, 14 Jul 2022

The massive loss of friction nighttime NEE data due to a high cutoff friction velocity has been seen at other sites, including Santarem km67 in Brazil, part of the LBA with a u* threshold of over 0.4 m/s. It's dubious to pick a differen threshold (0.3 in the authors' case) because the nighttime respiration is highly dependent on these data.
A novel correction has been proposed which may benefit these authors. The high dependence of NEE on u* at night may not be due to a "missing flux" but rather due to due storage pools that are "unseen" by the flux measurements and the storage profile. This may give a very different answer for the nighttime flux and overall carbon balance than using a u* threshold, which does not seem appropriate for the measurements at this site (or perhaps any site). https://doi.org/10.1016/j.agrformet.2017.12.186 https://harvardforest1.fas.harvard.edu/sites/harvardforest.fas.harvard.edu/files/publications/pdfs/Hayek_AgForMet_2018.pdf

Citation: https://doi.org/10.5194/egusphere-2022-384-CC1

Status: closed

RC1:
'Comment on egusphere-2022-384', Anonymous Referee #1, 27 Jun 2022

General comments

The question of the impact of source area heterogeneity on carbon (and other species) fluxes is highly relevant and interesting. Flux sites are often chosen where the heterogeneity is as small as possible but in reality, there are always some heterogeneity to consider. The title is however a bit misleading – it is not only heterogeneity of a forest but of a landscape. The first thought when reading such a title is that this must deal a lot with land cover and footprint analyses. But this is not the case. It is mainly a description of differences in fluxes in different wind sectors without any attempt to quantify differences in land cover and footprints in relation to these land cover variations and how this might impact the fluxes.

It is well known that the footprint area varies a lot depending on day/night unstable/stable conditions but this is hardly mentioned or only indirectly. Some of the figures, for instance Fig. 7e&f, where the difference in NEE between the two heights are plotted against u* and/or u for two different radiation levels gives a strong indication that stability has a large impact on the difference.

Another issue which might have impact on the fluxes are vertical and horizontal advection. It is not clear from the methods if there are concentration measurements at different heights in the tower but there are some mentioning of ‘plumbing’ which indicate that there might be such data available. If it is, then also total advection could be estimated. See for instance Yi et al Influence of advection on measurements of the net ecosystem-atmosphere exchange of CO2 from a very tall tower JGR Vol. 105, No. D8, 9991-9999, 2000.

The conclusion that it is the carbon enriched air from the clear cutting along the guy wires that are causing the spatial pattern of CO2 fluxes that are observed for the 30 m system is speculative without firm evidence. In the conclusion they also state that their hypothesis is confirmed and with the hypothesis that “the EC system located closer to the canopy will detect local features while the higher positioned system will detect naturally integrated CO2 flux”, this is obvious and no need to be ‘confirmed’.

Specific comments

Line 67-68: Describe the type of tower that is used. Please specify the width of the clearing along the guy wires. It is hardly visible in Fig. 2. You don’t mention if you have removed data from situations when the wind come through the tower. It would be good if you could indicate boom direction in some of your figures as well.

L 75: You present measurements over four years (which is good). Please state how often the gas analyzer was calibrated.

L 81: Here you mention that ‘soil efflux’ is measured with a transparent chamber. Later you refer to these measurements as ‘soil respiration’. If you are using a transparent chamber then you are measuring net exchange rather than respiration. Please clarify.

L 87: No measurement of net radiation? You don’t say anything about energy balance closure which is a common kind of quality measure for flux measurements.

L 106: Net ecosystem exchange NEE should include the storage term as well. You should make this clear if it was not and then name the flux otherwise.

L 155: Why not write out V/A instead of ‘h’ in eq. 4 to avoid the reader from reacting.

L 195 & Fig. 3. I don’t see that the wind profile is used at all anywhere in the paper. So you can remove it.

L294-296: Here it become important to know if you really measured ‘soil respiration’ or net flux from the forest floor.

L 375-376: There are papers on multi-level EC measurements from high towers. For instance the one mentioned above by Yi et al. and Davis et al. Global Change Biology (2003) 9, 1278–1293, Berger et al. Journal of Atmospheric and Oceanic Technology, Vol. 18, 529-542, 2001.

Fig. 2. It would be much more informative if the footprint climatology was divided in night & day and if more isolines were presented (e.g. 20,40, 60, 80%). And a map of land uses would also be more informative than an aerial photograph (which also can be shown for itself).

Fig. 11-12: here it would be really interesting with a proper footprint analyses.

Citation: https://doi.org/10.5194/egusphere-2022-384-RC1
- AC1: 'Reply on RC1', Alisa Krasnova, 12 Aug 2022
  
  We would like to thank the Anonymous Referee #1 for their time. Our answers are marked in bold
  General comments
  The question of the impact of source area heterogeneity on carbon (and other species) fluxes is highly relevant and interesting. Flux sites are often chosen where the heterogeneity is as small as possible but in reality, there are always some heterogeneity to consider. The title is however a bit misleading – it is not only heterogeneity of a forest but of a landscape. The first thought when reading such a title is that this must deal a lot with land cover and footprint analyses. But this is not the case. It is mainly a description of differences in fluxes in different wind sectors without any attempt to quantify differences in land cover and footprints in relation to these land cover variations and how this might impact the fluxes.
  Thank you for this comment! We will add footprint and land cover analysis to the revised manuscript. We also will consider changing the title.
  It is well known that the footprint area varies a lot depending on day/night unstable/stable conditions but this is hardly mentioned or only indirectly. Some of the figures, for instance Fig. 7e&f, where the difference in NEE between the two heights are plotted against u* and/or u for two different radiation levels gives a strong indication that stability has a large impact on the difference.
  Yes, you are correct. We will assess the impact of stability on the difference between the heights
  Another issue which might have impact on the fluxes are vertical and horizontal advection. It is not clear from the methods if there are concentration measurements at different heights in the tower but there are some mentioning of ‘plumbing’ which indicate that there might be such data available. If it is, then also total advection could be estimated. See for instance Yi et al Influence of advection on measurements of the net ecosystem-atmosphere exchange of CO2 from a very tall tower JGR Vol. 105, No. D8, 9991-9999, 2000.
  Thank you for the recommendation. We will add total advection estimation (we have concentration profile measurements at 30, 50 and 70 m)
  The conclusion that it is the carbon enriched air from the clear cutting along the guy wires that are causing the spatial pattern of CO2 fluxes that are observed for the 30 m system is speculative without firm evidence. In the conclusion they also state that their hypothesis is confirmed and with the hypothesis that “the EC system located closer to the canopy will detect local features while the higher positioned system will detect naturally integrated CO2 flux”, this is obvious and no need to be ‘confirmed’.
  Thank you, we will revise the conclusions
  Specific comments
  Line 67-68: Describe the type of tower that is used. Please specify the width of the clearing along the guy wires. It is hardly visible in Fig. 2. You don’t mention if you have removed data from situations when the wind come through the tower. It would be good if you could indicate boom direction in some of your figures as well.
  The information will be added to the revised manuscript
  L 75: You present measurements over four years (which is good). Please state how often the gas analyzer was calibrated.
  The information will be added to the revised manuscript
  L 81: Here you mention that ‘soil efflux’ is measured with a transparent chamber. Later you refer to these measurements as ‘soil respiration’. If you are using a transparent chamber then you are measuring net exchange rather than respiration. Please clarify.
  Yes, our soil chambers are transparent, meaning that the obtained fluxes represent soil floor respiration. After careful consideration we decided to omit soil fluxes from this study, due to their minor input to the final conclusions.
  L 87: No measurement of net radiation? You don’t say anything about energy balance closure which is a common kind of quality measure for flux measurements.
  The station is lacking soil heat flux measurements, so the energy balance closure estimation is not possible.
  L 106: Net ecosystem exchange NEE should include the storage term as well. You should make this clear if it was not and then name the flux otherwise.
  The information will be added to the revised manuscript
  L 155: Why not write out V/A instead of ‘h’ in eq. 4 to avoid the reader from reacting.
  Thank you for the suggestion. We decided to omit soil fluxes from this study, due to their minor input to the final conclusions.
  L 195 & Fig. 3. I don’t see that the wind profile is used at all anywhere in the paper. So you can remove it.
  Thank you, the figure will be removed
  L294-296: Here it become important to know if you really measured ‘soil respiration’ or net flux from the forest floor.
  Thank you for the suggestion. We decided to omit soil fluxes from this study, due to their minor input to the final conclusions.
  L 375-376: There are papers on multi-level EC measurements from high towers. For instance the one mentioned above by Yi et al. and Davis et al. Global Change Biology (2003) 9, 1278–1293, Berger et al. Journal of Atmospheric and Oceanic Technology, Vol. 18, 529-542, 2001.
  Thank you for the recommendation!
  Fig. 2. It would be much more informative if the footprint climatology was divided in night & day and if more isolines were presented (e.g. 20,40, 60, 80%). And a map of land uses would also be more informative than an aerial photograph (which also can be shown for itself).
  The map will be remade in the revised version of the manuscript
  Fig. 11-12: here it would be really interesting with a proper footprint analyses.
  The footprint analysis will be added to the revised manuscript
  
  Citation: https://doi.org/10.5194/egusphere-2022-384-AC1
RC2:
'Comment on egusphere-2022-384', Ankur Desai, 27 Jun 2022

Krasnova and colleagues evaluate sources of differences in eddy covariance fluxes made at two heights, 30 and 70 m on a tall tower at the SMEAR Estonia tower site. They conclude that the lower height is more affected by nearby clearings, leading to a smaller sink or larger sources, especially in calm conditions. Further, use of a wind directional rectification approach did not significantly alter this effect but allowed for estimation of a carbon budget in line with expectations. Overall, the contribution is useful for understanding the importance of evaluating sources of footprint bias and approaches to account for those in making reliable, defensible flux measurements over heterogenous ecosystems.

While the concept is useful, the manuscript lacks key consideration of drivers of flux profile in the atmospheric surface and boundary-layer. An eddy flux measured at a specific height (Fc) is not necessarily equivalent to the surface flux (NEE). Rather it is the flux at that height, and most flux profiles have linear slopes in the boundary-layer beyond the lower surface layer, of which 70 m would often not be part. So it should be no surprise that the two fluxes differ by some magnitude. Without accounting for flux divergence, advection, and storage flux, it is not clear that the two heights can be compared. My recommendation is thus the manuscript be rejected now, but a resubmission that addresses this oversight would be a positive contribution to the literature.

1.) In particular, the authors appear to have failed to account for the known effect of storage flux, the accumulation of flux under sensor height. Storage flux underestimation would be consistent with the delta_NEE (30-70) term being positive under a strong surface source, and during calm wind, low u* conditions at low light or night.

This is something I showed at our own 396 m tall tower site with fluxes measured at 30, 122, and 396 m for methane and CO2 fluxes in Desai et al., 2015. I encourage the authors to take a look and see if a similar analysis would be possible here using the CO2 concentration profile, even just from the two heights. Similarly, this particular site (Ameriflux/Fluxnet site US-PFa) has been operating since 1996 and there have been several publications looking at flux profiles and differences with height, including Davis et al., 2003, Berger et al., 2001, and Yi et al., 2000. It's too bad these were missed. Note, however, I do not require the authors to cite any work I mention - but hopefully these papers give you some guidance on how to approach the comparison here from a similar type of set up.

2.) I don't doubt that footprint differences are also a driving factor, but without an analysis of the flux profile and including storage flux (a known issues for tall towers back to foundational work on Massman and Lee, 2002), it is difficult to truly test your hypothesis. Beyond the storage flux issue, there is also the relatively qualitative way the authors try to evaluate the role of clearings and land surface with wind direction at the two heights. Since the authors have already run the Kljun 2D footprint model to construct the footprint climatology, I recommend developing a simple land cover map for both heights and evaluating the relative contributions to the footprint climatology. Also, given that the delta_NEE signal does have a wind directional signal, it would seem wise to try to look at the specific footprints during periods of max difference and identify whether the land cover fractions differ significantly (and whether the clearing is a major influence on the 30 m level).

I wonder if another way to test for this is to look at the same differences for CO2 but for H2O or heat flux? One might expect that lower GPP element from the clearing is driving the flux difference, one would also see differences in higher heat fluxes (more surface heating) and lower water fluxes (less transpiration). This is not a required correction, but could be a fun way to further support the hypothesis.

3.) The use of the chamber fluxes is relatively minimal. Personally, I find it provides compelling evidence that EC70 is more representative of the landscape (assuming the chambers are representative of the landscape), as both the magnitude and pattern more closely resemble EC70. I would suspect, given the forest cover here, that Rs would be more on the higher end of Rs/ER ratio report (0.3-0.8). Among other things the cited paper by Barba et al demonstrates that at many sites RSoil often exceeds REco, or is very similar in magnitude (look at Table 3 and the conclusion within it). It's not clear to me where the 0.3 to 0.8 range is actually mentioned in the citations. Davidson and Janssens makes no mention of ratios as far as I can tell. Either more work needs to be done to fully incorporate analysis of chamber fluxes, or it should be removed.

4.) The wind direction normalization from Hadden and Grelle (2017, also cited as 2016, is that a different paper?) seems like a good first test for estimating annual budgets, but there have been some more advanced approaches in recent years. For this manuscript, since the goal is just to demonstrate how consistent budgets are after normalization, probably not a major issues. But take a look at Griebel et al., 2020, Wang et al., 2006, Metzger, 2018.

Citations:

Berger, B. W., Davis, K. J., Yi, C., Bakwin, P. S., & Zhao, C. L. (2001). Long-Term Carbon Dioxide Fluxes from a Very Tall Tower in a Northern Forest: Flux Measurement Methodology, Journal of Atmospheric and Oceanic Technology, 18(4), 529-542. https://doi.org/10.1175/1520-0426(2001)018<0529:LTCDFF>2.0.CO;2

Davis, K.J., Bakwin, P.S., Yi, C., Berger, B.W., Zhao, C., Teclaw, R.M. and Isebrands, J.G. (2003), The annual cycles of CO2 and H2O exchange over a northern mixed forest as observed from a very tall tower. Global Change Biology, 9: 1278-1293. https://doi.org/10.1046/j.1365-2486.2003.00672.x

Desai, A.R., Xu, K., Tian H., Weishampel, P., Thom, J., Baumann, D., Andrews, A.E., Cook, B.D., King, J.Y., and Kolka, R., 2015. Landscape-level terrestrial methane flux observed from a very tall tower. Agricultural and Forest Meteorology, 201, 61-75, doi:10.1016/j.agrformet.2014.10.017.

Griebel, A., Metzen, D., Pendall, E., Burba, G., & Metzger, S. (2020). Generating spatially robust carbon budgets from flux tower observations. Geophysical Research Letters, 47, e2019GL085942. https://doi.org/10.1029/2019GL085942

Massman, W. J., & Lee, X. (2002). Eddy covariance flux corrections and uncertainties in long-term studies of carbon and energy exchanges. Agricultural and Forest Meteorology, 113(1), 121-144. doi:https://doi.org/10.1016/S0168-1923(02)00105-3

Metzger, 2018. Surface-atmosphere exchange in a box: Making the control volume a suitable representation for in-situ observations. AGricultural and Forest Meteorology, 255, 68-80, https://doi.org/10.1016/j.agrformet.2017.08.037

Wang, W., Davis, K. J., Cook, B. D., Butler, M. P., and Ricciuto, D. M. (2006), Decomposing CO2 fluxes measured over a mixed ecosystem at a tall tower and extending to a region: A case study, J. Geophys. Res., 111, G02005, doi:10.1029/2005JG000093.

Yi, C., Davis, K. J., Bakwin, P. S., Berger, B. W., and Marr, L. C. (2000), Influence of advection on measurements of the net ecosystem-atmosphere exchange of CO2 from a very tall tower, J. Geophys. Res., 105( D8), 9991– 9999, doi:10.1029/2000JD900080.

Sincerely,

Ankur Desai, Professor, UW-Madison

Citation: https://doi.org/10.5194/egusphere-2022-384-RC2
- AC2: 'Reply on RC2', Alisa Krasnova, 12 Aug 2022
  
  On behalf of all co-authors, I would like to thank you for the review and all the recommendations given. Please find our answers marked in bold.
  Krasnova and colleagues evaluate sources of differences in eddy covariance fluxes made at two heights, 30 and 70 m on a tall tower at the SMEAR Estonia tower site. They conclude that the lower height is more affected by nearby clearings, leading to a smaller sink or larger sources, especially in calm conditions. Further, use of a wind directional rectification approach did not significantly alter this effect but allowed for estimation of a carbon budget in line with expectations. Overall, the contribution is useful for understanding the importance of evaluating sources of footprint bias and approaches to account for those in making reliable, defensible flux measurements over heterogenous ecosystems.
  While the concept is useful, the manuscript lacks key consideration of drivers of flux profile in the atmospheric surface and boundary-layer. An eddy flux measured at a specific height (Fc) is not necessarily equivalent to the surface flux (NEE). Rather it is the flux at that height, and most flux profiles have linear slopes in the boundary-layer beyond the lower surface layer, of which 70 m would often not be part. So it should be no surprise that the two fluxes differ by some magnitude. Without accounting for flux divergence, advection, and storage flux, it is not clear that the two heights can be compared. My recommendation is thus the manuscript be rejected now, but a resubmission that addresses this oversight would be a positive contribution to the literature.
  Thank you for the valuable recommendation! We will revise the manuscript according to your suggestions.
  1.) In particular, the authors appear to have failed to account for the known effect of storage flux, the accumulation of flux under sensor height. Storage flux underestimation would be consistent with the delta_NEE (30-70) term being positive under a strong surface source, and during calm wind, low u* conditions at low light or night.
  In the absence of storage measurement system, the storage term was calculated with the simplified method using only concentrations at the respective heights. Unfortunately, this information was missing from the methods section. We will add it to the revised manuscript.
  This is something I showed at our own 396 m tall tower site with fluxes measured at 30, 122, and 396 m for methane and CO2 fluxes in Desai et al., 2015. I encourage the authors to take a look and see if a similar analysis would be possible here using the CO2 concentration profile, even just from the two heights. Similarly, this particular site (Ameriflux/Fluxnet site US-PFa) has been operating since 1996 and there have been several publications looking at flux profiles and differences with height, including Davis et al., 2003, Berger et al., 2001, and Yi et al., 2000. It's too bad these were missed. Note, however, I do not require the authors to cite any work I mention - but hopefully these papers give you some guidance on how to approach the comparison here from a similar type of set up.
  We appreciate your recommendations and will update our analysis for the revised version of the manuscript.
  2.) I don't doubt that footprint differences are also a driving factor, but without an analysis of the flux profile and including storage flux (a known issues for tall towers back to foundational work on Massman and Lee, 2002), it is difficult to truly test your hypothesis. Beyond the storage flux issue, there is also the relatively qualitative way the authors try to evaluate the role of clearings and land surface with wind direction at the two heights. Since the authors have already run the Kljun 2D footprint model to construct the footprint climatology, I recommend developing a simple land cover map for both heights and evaluating the relative contributions to the footprint climatology. Also, given that the delta_NEE signal does have a wind directional signal, it would seem wise to try to look at the specific footprints during periods of max difference and identify whether the land cover fractions differ significantly (and whether the clearing is a major influence on the 30 m level).
  Thank you for this suggestion! The footprint analysis will be added to the revised manuscript.
  I wonder if another way to test for this is to look at the same differences for CO2 but for H2O or heat flux? One might expect that lower GPP element from the clearing is driving the flux difference, one would also see differences in higher heat fluxes (more surface heating) and lower water fluxes (less transpiration). This is not a required correction, but could be a fun way to further support the hypothesis.
  Thank you, we will definitely try this!
  3.) The use of the chamber fluxes is relatively minimal. Personally, I find it provides compelling evidence that EC70 is more representative of the landscape (assuming the chambers are representative of the landscape), as both the magnitude and pattern more closely resemble EC70. I would suspect, given the forest cover here, that Rs would be more on the higher end of Rs/ER ratio report (0.3-0.8). Among other things the cited paper by Barba et al demonstrates that at many sites RSoil often exceeds REco, or is very similar in magnitude (look at Table 3 and the conclusion within it). It's not clear to me where the 0.3 to 0.8 range is actually mentioned in the citations. Davidson and Janssens makes no mention of ratios as far as I can tell. Either more work needs to be done to fully incorporate analysis of chamber fluxes, or it should be removed.
  After careful consideration we decided to omit soil fluxes from this study due to their minor input to the final conclusions.
  4.) The wind direction normalization from Hadden and Grelle (2017, also cited as 2016, is that a different paper?) seems like a good first test for estimating annual budgets, but there have been some more advanced approaches in recent years. For this manuscript, since the goal is just to demonstrate how consistent budgets are after normalization, probably not a major issues. But take a look at Griebel et al., 2020, Wang et al., 2006, Metzger, 2018.
  The simple approach by Hadden and Grelle (yes, they have 2 different papers, one in 2016 and another one is 2017) was applied to check if the observed differences were caused by the irregular distribution of wind direction. We will also look at the suggested publications.
  
  Citation: https://doi.org/10.5194/egusphere-2022-384-AC2
CC1: 'Comment on egusphere-2022-384', Matthew Hayek, 14 Jul 2022

The massive loss of friction nighttime NEE data due to a high cutoff friction velocity has been seen at other sites, including Santarem km67 in Brazil, part of the LBA with a u* threshold of over 0.4 m/s. It's dubious to pick a differen threshold (0.3 in the authors' case) because the nighttime respiration is highly dependent on these data.
A novel correction has been proposed which may benefit these authors. The high dependence of NEE on u* at night may not be due to a "missing flux" but rather due to due storage pools that are "unseen" by the flux measurements and the storage profile. This may give a very different answer for the nighttime flux and overall carbon balance than using a u* threshold, which does not seem appropriate for the measurements at this site (or perhaps any site). https://doi.org/10.1016/j.agrformet.2017.12.186 https://harvardforest1.fas.harvard.edu/sites/harvardforest.fas.harvard.edu/files/publications/pdfs/Hayek_AgForMet_2018.pdf

Citation: https://doi.org/10.5194/egusphere-2022-384-CC1

Alisa Krasnova, Dmitrii Krasnov, Hans Peter Ernst Cordey, and Steffen M. Noe

Viewed

Total article views: 830 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
475	331	24	830	11	16

HTML: 475
PDF: 331
XML: 24
Total: 830
BibTeX: 11
EndNote: 16

Views and downloads (calculated since 09 Jun 2022)

Month	HTML	PDF	XML	Total
Jun 2022	162	54	5	221
Jul 2022	43	11	1	55
Aug 2022	43	22	7	72
Sep 2022	9	13	0	22
Oct 2022	13	11	1	25
Nov 2022	12	11	0	23
Dec 2022	6	17	0	23
Jan 2023	8	9	0	17
Feb 2023	15	23	0	38
Mar 2023	12	15	0	27
Apr 2023	10	6	0	16
May 2023	5	8	0	13
Jun 2023	8	13	2	23
Jul 2023	4	12	2	18
Aug 2023	2	15	0	17
Sep 2023	12	15	0	27
Oct 2023	19	15	0	34
Nov 2023	10	5	0	15
Dec 2023	10	12	0	22
Jan 2024	21	16	2	39
Feb 2024	19	17	0	36
Mar 2024	23	8	3	34
Apr 2024	9	3	1	13

Cumulative views and downloads (calculated since 09 Jun 2022)

Month	HTML	PDF	XML	Total
Jun 2022	162	54	5	221
Jul 2022	43	11	1	55
Aug 2022	43	22	7	72
Sep 2022	9	13	0	22
Oct 2022	13	11	1	25
Nov 2022	12	11	0	23
Dec 2022	6	17	0	23
Jan 2023	8	9	0	17
Feb 2023	15	23	0	38
Mar 2023	12	15	0	27
Apr 2023	10	6	0	16
May 2023	5	8	0	13
Jun 2023	8	13	2	23
Jul 2023	4	12	2	18
Aug 2023	2	15	0	17
Sep 2023	12	15	0	27
Oct 2023	19	15	0	34
Nov 2023	10	5	0	15
Dec 2023	10	12	0	22
Jan 2024	21	16	2	39
Feb 2024	19	17	0	36
Mar 2024	23	8	3	34
Apr 2024	9	3	1	13

Viewed (geographical distribution)

Total article views: 771 (including HTML, PDF, and XML) Thereof 771 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 17 Apr 2024

Short summary

Forests play important role in global carbon cycle. Hemiboreal forests are located in a transitional zone between boreal and temperate zones and are characterised by higher variability in dominating tree species. We analysed carbon exchange of a hemiboreal forest using two measurement systems located on the same tower. The lower system could detect the influence of clearcuts and tower guywire tunnels, while the higher system allowed to assess the carbon balance of the full landscape.


Total:	0
HTML:	0
PDF:	0
XML:	0