How atmospheric CO<sub>2</sub> can inform us on annual and decadal shifts in the biospheric carbon uptake period

Kariyathan, Theertha; Bastos, Ana; Reichstein, Markus; Peters, Wouter; Marshall, Julia

doi:https://doi.org/10.5194/egusphere-2024-1382

Preprints

https://doi.org/10.5194/egusphere-2024-1382

Preprints

05 Jul 2024

| 05 Jul 2024

How atmospheric CO₂ can inform us on annual and decadal shifts in the biospheric carbon uptake period

Theertha Kariyathan, Ana Bastos, Markus Reichstein, Wouter Peters, and Julia Marshall

Abstract. The carbon uptake period (CUP) refers to the time of each year during which the rate of photosynthetic uptake surpasses that of respiration in the terrestrial biosphere, resulting in a net absorption of CO₂ from the atmosphere to the land. Since climate drivers influence both photosynthesis and respiration, the CUP offers valuable insights into how the terrestrial biosphere responds to climate variations and affects the carbon budget. Several studies have assessed large-scale changes in CUP based on seasonal metrics from CO₂ mole fraction measurements. However, an in-depth understanding of the sensitivity of the CUP as derived from the CO₂ mole fraction data (CUP_MR) to actual changes in the CUP of the net ecosystem exchange (CUP_NEE) is missing. In this study, we specifically assess the impact of (i) atmospheric transport (ii) inter-annual variability in CUP_NEE (iii) regional contribution to the signals that integrate at different background sites where CO₂ dry air mole fraction measurements are made. We conducted idealized simulations where we imposed known changes (∆) to the CUP_NEE in the Northern Hemisphere to test the effect of the aforementioned factors in CUP_MR metrics at ten Northern Hemisphere sites. Our analysis indicates a significant damping of changes in the simulated ∆CUP_MR due to the integration of signals with varying CUP_NEE timing across regions. CUP_MR at well-studied sites such as Mauna Loa, Barrow, and Alert showed only 50 % of the applied ∆CUP_NEE under non interannually-varying atmospheric transport conditions. Further, our synthetic analyses conclude that interannual variability (IAV) in atmospheric transport accounts for a significant part of the changes in the observed signals. However, even after separating the contribution of transport IAV, the estimates of surface changes in CUP by previous studies are not likely to provide an accurate magnitude of the actual changes occurring over the surface. The observed signal experiences significant damping as the atmosphere averages out non-synchronous signals from various regions.

Received: 10 May 2024 – Discussion started: 05 Jul 2024

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.

Download & links

Preprint (PDF, 4489 KB)

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 (4489 KB)

Download & links

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Journal article(s) based on this preprint

24 Jul 2025

Limitations in the use of atmospheric CO₂ observations to directly infer changes in the length of the biospheric carbon uptake period

Theertha Kariyathan, Ana Bastos, Markus Reichstein, Wouter Peters, and Julia Marshall

Atmos. Chem. Phys., 25, 7863–7878, https://doi.org/10.5194/acp-25-7863-2025,https://doi.org/10.5194/acp-25-7863-2025, 2025

Short summary

Theertha Kariyathan, Ana Bastos, Markus Reichstein, Wouter Peters, and Julia Marshall

Interactive discussion

Status: closed

CC1:
'The meaning of mixing ratio', Andrew Kowalski, 10 Jul 2024

The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-1382/egusphere-2024-1382-CC1-supplement.pdf

Citation: https://doi.org/10.5194/egusphere-2024-1382-CC1
- AC1: 'Reply on CC1', Theertha Kariyathan, 15 Jul 2024
  
  Reply on RCC1
  
  Thank you for your insightful comments. We appreciate the opportunity to clarify the definitions and terminologies used in our manuscript.
  
  The reviewer is correct that atmospheric CO₂ measurements are reported in dry air mole fractions, excluding water vapor. And we agree that we could have more explicitly added this to the wording we use throughout the manuscript.
  
  The TM3 model operates using the mass of dry air and the mass of specified tracers, which are then recalculated to mole fractions based on the tracer molar mass and the molar mass of dry air. So while the actual atmosphere contains both dry air and water vapor, the tracer transport modeling approach works with an atmosphere that is entirely "dry" mass (Heimann and Körner, 2003). Although this sounds like a poor approach, this is only a small source of potential errors for offline transport models, as we will explain.
  
  In the TM3 model, the atmospheric mass is initialized as dry mass from surface pressure fields derived from the parent weather model, the Integrated Forecast System (IFS). In this conversion of surface pressure to mass we disregard the pressure contribution from lighter water vapor. The CO₂ mass is determined by multiplying the air mass by the dry air mole fraction (approximately 400 ppm) and the ratio of CO₂ molar mass (44 g/mol) to dry air molar mass (28.96 g/mol). The CO₂ sources and sinks are then modeled by adjusting the CO₂ mass accordingly, and finally converting the updated mass back to a CO₂ mole fraction using the same dry air assumption (28.96/44). This consistent exclusion of water vapor mass ensures that the model results are directly comparable to the measured dry air mole fractions, and also maintains full mass balance in the CO₂ budget independent of water vapor variations. Finally, the inclusion of water vapor differences between the time of introducing sources and sinks, and recording mole fractions is considered minimal due to the small variation in water vapor mole fraction (<1%) that then only affects the simulated tracer change (few ppm). For example, Lee and Weidner (2016) simulated CO₂ fluxes using the GEOS-Chem Adjoint (GCA) system under two conditions: 1) assuming a dry pressure surface and 2) assuming a wet pressure surface and found a bias in the global CO₂ volume mixing ratio of less than 0.1%. This result was independently confirmed using the TM5 model (unpublished), which shares a similar modeling framework with the TM3 model. This effect is also much smaller than the magnitude of changes that we aim to detect and their much larger uncertainty from interannual variability in transport (see Kariyathan et al., 2023).
  
  The reviewer also inquired whether including water vapor mass (and its gradients) would affect the transport modeling. For advection, the mass fluxes in our offline model are based on pressure gradients derived from the Integrated Forecast System (IFS), which includes water vapor in its calculations and hence there is no reason to assume these mass-fluxes are incorrect. For turbulent motions in the planetary boundary layer, water vapor is considered when setting up the vertical diffusion constants (K), but it is not included in the calculations for molecular diffusion or Steffen flow at the leaf level. However, in large models with grid resolutions of 100-300 km and for tracers at small ambient levels, these will play a small role.
  
  Reference
  
  Heimann, M. and Körner, S., 2003, The Global Atmospheric Tracer Model TM3: Model Description and User’s Manual, Release 3.8a, Max-Planck-Institut für Biogeochemie, Jena, Germany, 131pp.
  
  Kariyathan, T., Bastos, A., Reichstein, M., Peters, W., and Marshall, J., 2024, How atmospheric CO₂ can inform us on annual and decadal shifts in the biospheric carbon uptake period, EGUsphere, 2024, 1-22, https://egusphere.copernicus.org/preprints/2024/egusphere-2024-1382, DOI:10.5194/egusphere-2024-1382
  
  Lee, M. , and Weidner, R. (2016). Jpl publication 16‐4: Surface pressure dependencies in the GEOS‐chem adjoint system and the impact of the GEOS‐5 surface pressure on CO₂ model forecast. Jet Propulsion Laboratory California Institute of Technology Pasadena, California.
  
  Citation: https://doi.org/10.5194/egusphere-2024-1382-AC1
RC1:
'Comment on egusphere-2024-1382', Anonymous Referee #2, 14 Sep 2024
Review of "How atmospheric CO₂ can inform us on annual and decadal shifts in the biospheric carbon uptake period" by Kariyathan et al.
The authors have conducted a sensitivity analysis of the carbon uptake period (CUP) in the atmospheric CO₂ mole fraction to understand how changes in the CUP in net ecosystem exchange (NEE) are reflected in the observed atmospheric mole fraction. Using a series of model simulations, they examined how atmospheric transport and interannual variations in NEE contribute to changes in the CUP in the atmospheric CO₂ mole fraction, focusing on 10 observations sites across the northern hemisphere. The authors showed that interannual variations in atmospheric transport provides a significant contribution to changes in the mole fraction CUP. They found that changes in NEE are manifested as damped variations in the CO₂ mole fraction, suggesting that the use of observations of changes in the mole fraction CUP to quantify changes in NEE will likely result in an underestimate of the changes in NEE. The authors have developed an interesting approach for assessing the impact of changes in NEE on the atmospheric CO₂ mole fraction that would be a valuable contribution to the literature. However, as discussed in my comments below, these results are not surprising. I have some concerns about the observation sites that were selected for the analysis and the resulting conclusions. I would like to see the authors address these concerns before the manuscript can be considered suitable for publication.
Main comments
The authors are using 10 observation sites in the northern hemisphere to assess the sensitivity of changes in the CO₂ mole fraction to variations in NEE. However, the sites that were selected are those that capture mainly background CO₂. As the authors noted, the CO₂ model fraction at these background sites "represent the balance between surface emissions and uptake from land and ocean… over large spatial scales," so it is not surprising that these sites will capture a damped manifestation of the changes in NEE. At the end of Section 4.2, the authors stated that "CO₂ observations should preferably be interpreted following a formal inverse estimate of the corresponding surface NEE. It is then possible to account for the inter-annual variability, trends and delays imposed by the slow atmospheric mixing." However, inversions using the remote background sites also have difficulties in providing robust estimates of regional changes in NEE because of the influence of transport and mixing. This is one of the main reasons there has been significant effort focused on expanding the observing network. It is reasonable to ask how would the results of the study change if different sites were selected? For example, would Hegyhatsal (HUN) in Hungary be more sensitive to variations in changes in European NEE? Would East Trout Lake (ETL) in Canada be more sensitive to North American boreal fluxes? There is clearly a need to avoid sites that are strongly sensitive to anthropogenic emissions, but are the 10 sites used in this study the best sites for capturing changes in NEE given the well-known confounding influence of transport and mixing? As a modeling study, it seems to me that the authors have the opportunity to better evaluate the sensitivity of the existing observing network rather than that of what seems to be an arbitrary selection of 10 sites in the network.

Another issue is that the authors are using the temporal frequency of the flask measurements, but that is quite coarse – it is at best weekly. Some sites provide continuous measurements. Is there any benefit to using continuous data for capturing the variations and trends in NEE?

The analysis uses the ensemble of the first derivative (EFD) method to estimate the CUP. This approach was described in Kariyathan et al. (2023), but there is no explanation of the method in this manuscript. The reader is forced to go through Kariyathan et al. (2023) to understand what is being done. A brief description of the approach in the current manuscript would be helpful to the reader.

Other comments
Table 1 caption: I believe the subscript refers to the CUP_NEE, so the caption should read "subscript and superscript of the last character denote CUP_NEE and variability (V) in transport, respectively."

Line 90: Please change: "we used experiment ENV₀⁰ and LNV" to "we used experiments ENV₀⁰ and LNV."

Line 100: Please list the regions over which the fluxes were aggregated.

Lines 137-143: Should Figure 3 be referenced in the text here? I don't believe the figure is actually referenced in the manuscript.

Figure 4 caption: On the second line, change "left panel shows" to "left panels show". On the third line form the bottom, change "right panel ((d) to (f)) shows" to "right panels ((d) to (f)) show".

Line 218: It would be helpful to see what are the results for the other sites. Can this be presented in a separate table?

Lines 226-227: It is unclear that it is the trend from the IAV in transport that "leads to the asymmetry". Perhaps this should be phrased as "the trend from the IAV in transport contributes to the asymmetry."

Line 300: This description is confusing. Transport from Europe, Russia, and boreal Eurasia is westerly, so it is confusing when the authors say that "air mass transport is largely over the Northern Atlantic Ocean and does not extend into the continents in some years." During summer, because of the warm surface it is difficult for air transported from Europe and boreal Eurasia to enter the Arctic at low altitudes, so this circumpolar westerly transport occurs at lower latitudes, reducing the sensitivity of the high-latitude ZEP site to Eurasian air masses. I believe that the authors are trying to say that because of the lower latitude circumpolar transport in summer, the sensitivity of the ZEP site to variations in surface CO₂ fluxes in summer is confined to the Arctic and does not extend equatorward into the continents.
Citation: https://doi.org/10.5194/egusphere-2024-1382-RC1
RC2:
'Comment on egusphere-2024-1382', Anonymous Referee #3, 19 Oct 2024
Summary:
The authors present a modeling study to determine how much surface flux information is contained in the atmospheric CO2 mole fraction observations. They see how well they can recover the time duration when net ecosystem exchange is uptake by the northern hemisphere biosphere from the seasonal cycle of atmospheric CO2 mole fraction observations after running surface fluxes and anomalies through an atmospheric transport model. They show that the surface flux signal information is reduced by mixing in the atmosphere.
Major comments:
The authors motivate this study by a few published studies of changes in the atmospheric CO2 mole fraction seasonal cycle used to infer northern hemisphere surface flux changes without explicitly considering atmospheric transport. This feels like a strawman. The community already acknowledges the influence of atmospheric transport when linking concentrations to surface fluxes. That is why atmospheric inverse techniques were developed decades ago. Jin et al. (2022) provides an example of a study that uses an atmospheric transport model to separate the influence of surface fluxes and winds on the seasonal cycle at Mauna Loa without using an inverse approach. The references therein identify other studies that also make the point that atmospheric CO2 mole fractions are influenced by fluxes and winds. The authors’ most important sentence buried in the middle of the discussion is the main point of this study. “Therefore, CO2 observations should preferably be interpreted following a formal inversion estimate of the corresponding surface NEE. It is then possible to account for the inter-annual variability, trends and delays imposed by the slow atmospheric mixing.” I’m not sure what new perspective this study adds.

Furthermore, the way that the wind influence is described as “the integration of signals with varying CUPNEE timing across regions” and “variations in the timing of CUPNEE across different regions from where the signal is integrated” and in Fig 9 is confusing. They are describing the influence of atmospheric transport or winds, but making it sound like it’s a flux synchronization issue.

Other studies have mostly focused on CO2 seasonal cycle amplitude changes and this study’s focus on the duration of the carbon uptake period is a little unique. But this study does hypothetical forward model experiments without using the observed atmospheric CO2 mole fractions to constrain which fluxes anomalies are supported by the observations, if any. What new information about the behavior of the northern hemisphere biosphere is learned here? The authors state “Considering the {transport} reducing factor, a change of approximately 0.7 days/year might be occurring in the surface fluxes.”

The nomenclature for the experiments makes it difficult to look at a figure and quickly interpret which experiment combination of fluxes and winds it is comparing. The reader must work very hard to understand and remember the notation. Table 1 helps, but does it have to be that complicated? It seems the paired notation of the Experiment and Delta CUPNEE columns are needed to uniquely identify the tests, especially with ENVTo and LNVTo cases.

Misleading title? The finding was that obs can’t inform well on surface fluxes, without considering atmospheric transport.

References:
Jin, Y., Keeling, R. F., Rödenbeck, C., Patra, P. K., Piper, S. C., & Schwartzman, A. (2022). Impact of changing winds on the Mauna Loa CO₂ seasonal cycle in relation to the Pacific Decadal Oscillation. Journal of Geophysical Research: Atmospheres, 127, e2021JD035892. https://doi.org/10.1029/2021JD035892
Citation: https://doi.org/10.5194/egusphere-2024-1382-RC2

Interactive discussion

Status: closed

CC1:
'The meaning of mixing ratio', Andrew Kowalski, 10 Jul 2024

The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-1382/egusphere-2024-1382-CC1-supplement.pdf

Citation: https://doi.org/10.5194/egusphere-2024-1382-CC1
- AC1: 'Reply on CC1', Theertha Kariyathan, 15 Jul 2024
  
  Reply on RCC1
  
  Thank you for your insightful comments. We appreciate the opportunity to clarify the definitions and terminologies used in our manuscript.
  
  The reviewer is correct that atmospheric CO₂ measurements are reported in dry air mole fractions, excluding water vapor. And we agree that we could have more explicitly added this to the wording we use throughout the manuscript.
  
  The TM3 model operates using the mass of dry air and the mass of specified tracers, which are then recalculated to mole fractions based on the tracer molar mass and the molar mass of dry air. So while the actual atmosphere contains both dry air and water vapor, the tracer transport modeling approach works with an atmosphere that is entirely "dry" mass (Heimann and Körner, 2003). Although this sounds like a poor approach, this is only a small source of potential errors for offline transport models, as we will explain.
  
  In the TM3 model, the atmospheric mass is initialized as dry mass from surface pressure fields derived from the parent weather model, the Integrated Forecast System (IFS). In this conversion of surface pressure to mass we disregard the pressure contribution from lighter water vapor. The CO₂ mass is determined by multiplying the air mass by the dry air mole fraction (approximately 400 ppm) and the ratio of CO₂ molar mass (44 g/mol) to dry air molar mass (28.96 g/mol). The CO₂ sources and sinks are then modeled by adjusting the CO₂ mass accordingly, and finally converting the updated mass back to a CO₂ mole fraction using the same dry air assumption (28.96/44). This consistent exclusion of water vapor mass ensures that the model results are directly comparable to the measured dry air mole fractions, and also maintains full mass balance in the CO₂ budget independent of water vapor variations. Finally, the inclusion of water vapor differences between the time of introducing sources and sinks, and recording mole fractions is considered minimal due to the small variation in water vapor mole fraction (<1%) that then only affects the simulated tracer change (few ppm). For example, Lee and Weidner (2016) simulated CO₂ fluxes using the GEOS-Chem Adjoint (GCA) system under two conditions: 1) assuming a dry pressure surface and 2) assuming a wet pressure surface and found a bias in the global CO₂ volume mixing ratio of less than 0.1%. This result was independently confirmed using the TM5 model (unpublished), which shares a similar modeling framework with the TM3 model. This effect is also much smaller than the magnitude of changes that we aim to detect and their much larger uncertainty from interannual variability in transport (see Kariyathan et al., 2023).
  
  The reviewer also inquired whether including water vapor mass (and its gradients) would affect the transport modeling. For advection, the mass fluxes in our offline model are based on pressure gradients derived from the Integrated Forecast System (IFS), which includes water vapor in its calculations and hence there is no reason to assume these mass-fluxes are incorrect. For turbulent motions in the planetary boundary layer, water vapor is considered when setting up the vertical diffusion constants (K), but it is not included in the calculations for molecular diffusion or Steffen flow at the leaf level. However, in large models with grid resolutions of 100-300 km and for tracers at small ambient levels, these will play a small role.
  
  Reference
  
  Heimann, M. and Körner, S., 2003, The Global Atmospheric Tracer Model TM3: Model Description and User’s Manual, Release 3.8a, Max-Planck-Institut für Biogeochemie, Jena, Germany, 131pp.
  
  Kariyathan, T., Bastos, A., Reichstein, M., Peters, W., and Marshall, J., 2024, How atmospheric CO₂ can inform us on annual and decadal shifts in the biospheric carbon uptake period, EGUsphere, 2024, 1-22, https://egusphere.copernicus.org/preprints/2024/egusphere-2024-1382, DOI:10.5194/egusphere-2024-1382
  
  Lee, M. , and Weidner, R. (2016). Jpl publication 16‐4: Surface pressure dependencies in the GEOS‐chem adjoint system and the impact of the GEOS‐5 surface pressure on CO₂ model forecast. Jet Propulsion Laboratory California Institute of Technology Pasadena, California.
  
  Citation: https://doi.org/10.5194/egusphere-2024-1382-AC1
RC1:
'Comment on egusphere-2024-1382', Anonymous Referee #2, 14 Sep 2024
Review of "How atmospheric CO₂ can inform us on annual and decadal shifts in the biospheric carbon uptake period" by Kariyathan et al.
The authors have conducted a sensitivity analysis of the carbon uptake period (CUP) in the atmospheric CO₂ mole fraction to understand how changes in the CUP in net ecosystem exchange (NEE) are reflected in the observed atmospheric mole fraction. Using a series of model simulations, they examined how atmospheric transport and interannual variations in NEE contribute to changes in the CUP in the atmospheric CO₂ mole fraction, focusing on 10 observations sites across the northern hemisphere. The authors showed that interannual variations in atmospheric transport provides a significant contribution to changes in the mole fraction CUP. They found that changes in NEE are manifested as damped variations in the CO₂ mole fraction, suggesting that the use of observations of changes in the mole fraction CUP to quantify changes in NEE will likely result in an underestimate of the changes in NEE. The authors have developed an interesting approach for assessing the impact of changes in NEE on the atmospheric CO₂ mole fraction that would be a valuable contribution to the literature. However, as discussed in my comments below, these results are not surprising. I have some concerns about the observation sites that were selected for the analysis and the resulting conclusions. I would like to see the authors address these concerns before the manuscript can be considered suitable for publication.
Main comments
The authors are using 10 observation sites in the northern hemisphere to assess the sensitivity of changes in the CO₂ mole fraction to variations in NEE. However, the sites that were selected are those that capture mainly background CO₂. As the authors noted, the CO₂ model fraction at these background sites "represent the balance between surface emissions and uptake from land and ocean… over large spatial scales," so it is not surprising that these sites will capture a damped manifestation of the changes in NEE. At the end of Section 4.2, the authors stated that "CO₂ observations should preferably be interpreted following a formal inverse estimate of the corresponding surface NEE. It is then possible to account for the inter-annual variability, trends and delays imposed by the slow atmospheric mixing." However, inversions using the remote background sites also have difficulties in providing robust estimates of regional changes in NEE because of the influence of transport and mixing. This is one of the main reasons there has been significant effort focused on expanding the observing network. It is reasonable to ask how would the results of the study change if different sites were selected? For example, would Hegyhatsal (HUN) in Hungary be more sensitive to variations in changes in European NEE? Would East Trout Lake (ETL) in Canada be more sensitive to North American boreal fluxes? There is clearly a need to avoid sites that are strongly sensitive to anthropogenic emissions, but are the 10 sites used in this study the best sites for capturing changes in NEE given the well-known confounding influence of transport and mixing? As a modeling study, it seems to me that the authors have the opportunity to better evaluate the sensitivity of the existing observing network rather than that of what seems to be an arbitrary selection of 10 sites in the network.

Another issue is that the authors are using the temporal frequency of the flask measurements, but that is quite coarse – it is at best weekly. Some sites provide continuous measurements. Is there any benefit to using continuous data for capturing the variations and trends in NEE?

The analysis uses the ensemble of the first derivative (EFD) method to estimate the CUP. This approach was described in Kariyathan et al. (2023), but there is no explanation of the method in this manuscript. The reader is forced to go through Kariyathan et al. (2023) to understand what is being done. A brief description of the approach in the current manuscript would be helpful to the reader.

Other comments
Table 1 caption: I believe the subscript refers to the CUP_NEE, so the caption should read "subscript and superscript of the last character denote CUP_NEE and variability (V) in transport, respectively."

Line 90: Please change: "we used experiment ENV₀⁰ and LNV" to "we used experiments ENV₀⁰ and LNV."

Line 100: Please list the regions over which the fluxes were aggregated.

Lines 137-143: Should Figure 3 be referenced in the text here? I don't believe the figure is actually referenced in the manuscript.

Figure 4 caption: On the second line, change "left panel shows" to "left panels show". On the third line form the bottom, change "right panel ((d) to (f)) shows" to "right panels ((d) to (f)) show".

Line 218: It would be helpful to see what are the results for the other sites. Can this be presented in a separate table?

Lines 226-227: It is unclear that it is the trend from the IAV in transport that "leads to the asymmetry". Perhaps this should be phrased as "the trend from the IAV in transport contributes to the asymmetry."

Line 300: This description is confusing. Transport from Europe, Russia, and boreal Eurasia is westerly, so it is confusing when the authors say that "air mass transport is largely over the Northern Atlantic Ocean and does not extend into the continents in some years." During summer, because of the warm surface it is difficult for air transported from Europe and boreal Eurasia to enter the Arctic at low altitudes, so this circumpolar westerly transport occurs at lower latitudes, reducing the sensitivity of the high-latitude ZEP site to Eurasian air masses. I believe that the authors are trying to say that because of the lower latitude circumpolar transport in summer, the sensitivity of the ZEP site to variations in surface CO₂ fluxes in summer is confined to the Arctic and does not extend equatorward into the continents.
Citation: https://doi.org/10.5194/egusphere-2024-1382-RC1
RC2:
'Comment on egusphere-2024-1382', Anonymous Referee #3, 19 Oct 2024
Summary:
The authors present a modeling study to determine how much surface flux information is contained in the atmospheric CO2 mole fraction observations. They see how well they can recover the time duration when net ecosystem exchange is uptake by the northern hemisphere biosphere from the seasonal cycle of atmospheric CO2 mole fraction observations after running surface fluxes and anomalies through an atmospheric transport model. They show that the surface flux signal information is reduced by mixing in the atmosphere.
Major comments:
The authors motivate this study by a few published studies of changes in the atmospheric CO2 mole fraction seasonal cycle used to infer northern hemisphere surface flux changes without explicitly considering atmospheric transport. This feels like a strawman. The community already acknowledges the influence of atmospheric transport when linking concentrations to surface fluxes. That is why atmospheric inverse techniques were developed decades ago. Jin et al. (2022) provides an example of a study that uses an atmospheric transport model to separate the influence of surface fluxes and winds on the seasonal cycle at Mauna Loa without using an inverse approach. The references therein identify other studies that also make the point that atmospheric CO2 mole fractions are influenced by fluxes and winds. The authors’ most important sentence buried in the middle of the discussion is the main point of this study. “Therefore, CO2 observations should preferably be interpreted following a formal inversion estimate of the corresponding surface NEE. It is then possible to account for the inter-annual variability, trends and delays imposed by the slow atmospheric mixing.” I’m not sure what new perspective this study adds.

Furthermore, the way that the wind influence is described as “the integration of signals with varying CUPNEE timing across regions” and “variations in the timing of CUPNEE across different regions from where the signal is integrated” and in Fig 9 is confusing. They are describing the influence of atmospheric transport or winds, but making it sound like it’s a flux synchronization issue.

Other studies have mostly focused on CO2 seasonal cycle amplitude changes and this study’s focus on the duration of the carbon uptake period is a little unique. But this study does hypothetical forward model experiments without using the observed atmospheric CO2 mole fractions to constrain which fluxes anomalies are supported by the observations, if any. What new information about the behavior of the northern hemisphere biosphere is learned here? The authors state “Considering the {transport} reducing factor, a change of approximately 0.7 days/year might be occurring in the surface fluxes.”

The nomenclature for the experiments makes it difficult to look at a figure and quickly interpret which experiment combination of fluxes and winds it is comparing. The reader must work very hard to understand and remember the notation. Table 1 helps, but does it have to be that complicated? It seems the paired notation of the Experiment and Delta CUPNEE columns are needed to uniquely identify the tests, especially with ENVTo and LNVTo cases.

Misleading title? The finding was that obs can’t inform well on surface fluxes, without considering atmospheric transport.

References:
Jin, Y., Keeling, R. F., Rödenbeck, C., Patra, P. K., Piper, S. C., & Schwartzman, A. (2022). Impact of changing winds on the Mauna Loa CO₂ seasonal cycle in relation to the Pacific Decadal Oscillation. Journal of Geophysical Research: Atmospheres, 127, e2021JD035892. https://doi.org/10.1029/2021JD035892
Citation: https://doi.org/10.5194/egusphere-2024-1382-RC2

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision | EF: Editorial file upload

AR by Theertha Kariyathan on behalf of the Authors (16 Dec 2024) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (02 Jan 2025) by Amos Tai

RR by Anonymous Referee #2 (13 Feb 2025)

RR by Anonymous Referee #3 (16 Feb 2025)

Suggestions for revision or reasons for rejection

Main issues:
1) Flux asynchronization versus transport lags from one region to another
The authors responded to my previous comment aiming to draw a distinction between the lag time that it takes flux anomalies from one region to appear in the atmosphere above a downwind region and the idea of flux asynchronization, for example, that the downward region starts taking up CO2 later than the upwind region. They authors use the results in Figure 9 to infer that the start of the start of the CUP is roughly a month later in eastern Eurasia compared to western Eurasia. If this is true, the authors can verify this with a map of CUP start days. Take this thought experiment of mixing ratio anomalies in the constant delay of CUP_NEE experiment. In May, the delayed start of spring leads to anomalous high CO2 mixing ratios over Eurasia. These anomalies would be strongest in places with the largest absolute (not relative) change in CO2 fluxes. At the next month timestep, anomalies of high CO2 over western Eurasia from a change in May NEE are transported to Eastern Eurasia and combined with June NEE anomalies in Eastern Eurasia, leading to a larger magnitude of CO2 anomalies. By August, the decrease in total CO2 uptake in Eurasia is well mixed through the northern hemisphere. That doesn’t require differences in the timing of CO2 uptake from place-to-place. It is possible that zonal mixing is far too fast for this transport lag to make sense. This might depend on the time step of the atmospheric transport model. If so, the map of CUP start days should confirm the authors’ interpretation of these signals and could be included as a supplemental figure.
Related, I am struggling to understand the reason for the results in Figure 10. If a 10-day advance is applied everywhere, how is the ‘true’ value not identical to the uniform forcing? “Due to the difference in CUPNEE timing across the pixels, the “true” ΔCUPNEE of a region for a period will be different from the applied Δ, as shown in Fig. 10. In a significant portion (54% pixels) of the North American Boreal region (Fig. 10), the onset of CUPNEE typically occurs between days 110 and 150. If CUPNEE fluxes are integrated during this period, the observed signal will represent a substantial portion of the CUPNEE onset from the region. However, when a 10-day advancement in CUPNEE onset is applied, the true ΔCUPNEE of the region during this period is only 8.9 days (average of ΔCUPNEE over the region shown in Fig. 10) and the CUPMR change at the observation site would be less than the applied 10-day Δ.” Does this mean that some pixels are a net source of CO2 while other pixels have a 10-day advancement in CO2 uptake. Therefore, the model is averaging pixels with a 10 day advancement and ones with zero day advancement because the CUP hasn’t happened in those locations yet? For this to be true, the pixel of CUP changes must be smaller than the grid cell in Fig 10?

2) Value of the study
The authors made minor edits that highlight that this study is unique from other studies on the influence of atmospheric transport on atmospheric CO2 variability is in the days of net carbon uptake, rather than zero crossing days or seasonal amplitude. The authors should edit the conclusions to remind the reader that these conclusions are based on the forward model experiments and do not inform on actual biospheric changes or derived from atmospheric observations.

3) Specific comments

Do all the studies cited in this passage actually use CUP_MR metrics? “A longer growing season does not necessarily mean an increase in CUPNEE or CUPMR as they are determined by both photosynthesis and respiration. The existing literature on the CUPNEE changes in the Northern Hemisphere based on CUPMR varies from increasing (Murayama et al., 2007) to neutral (Barichivich et al., 2012) to decreasing (Piao et al., 2008; Fu et al., 2017).”

This conclusion is typical of the signal-to-noise problem that challenges any trend detection efforts. “While sites like BRW and SHM captured the prescribed long-term changes, they proved insensitive when IAV in CUPNEE was doubled.”

What does the last sentence mean? Particularly the last half? “Furthermore, trends prescribed to individual TransCom3 regions were not captured by the evaluated sites, showing the long-term changes in the seasonal cycle of time series.”

Hide

ED: Publish subject to minor revisions (review by editor) (25 Feb 2025) by Amos Tai

AR by Theertha Kariyathan on behalf of the Authors (20 Mar 2025) Author's response Author's tracked changes Manuscript

ED: Publish as is (10 Apr 2025) by Amos Tai

AR by Theertha Kariyathan on behalf of the Authors (17 Apr 2025)

Journal article(s) based on this preprint

24 Jul 2025

Limitations in the use of atmospheric CO₂ observations to directly infer changes in the length of the biospheric carbon uptake period

Theertha Kariyathan, Ana Bastos, Markus Reichstein, Wouter Peters, and Julia Marshall

Atmos. Chem. Phys., 25, 7863–7878, https://doi.org/10.5194/acp-25-7863-2025,https://doi.org/10.5194/acp-25-7863-2025, 2025

Short summary

Theertha Kariyathan, Ana Bastos, Markus Reichstein, Wouter Peters, and Julia Marshall

Viewed

Total article views: 904 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
626	162	116	904	39	46

HTML: 626
PDF: 162
XML: 116
Total: 904
BibTeX: 39
EndNote: 46

Views and downloads (calculated since 05 Jul 2024)

Month	HTML	PDF	XML	Total
Jul 2024	256	89	20	365
Aug 2024	42	8	5	55
Sep 2024	53	7	3	63
Oct 2024	57	14	1	72
Nov 2024	39	1	1	41
Dec 2024	25	3	0	28
Jan 2025	11	5	6	22
Feb 2025	11	2	7	20
Mar 2025	14	8	2	24
Apr 2025	22	11	18	51
May 2025	24	5	44	73
Jun 2025	49	4	8	61
Jul 2025	23	5	1	29

Cumulative views and downloads (calculated since 05 Jul 2024)

Month	HTML	PDF	XML	Total
Jul 2024	256	89	20	365
Aug 2024	42	8	5	55
Sep 2024	53	7	3	63
Oct 2024	57	14	1	72
Nov 2024	39	1	1	41
Dec 2024	25	3	0	28
Jan 2025	11	5	6	22
Feb 2025	11	2	7	20
Mar 2025	14	8	2	24
Apr 2025	22	11	18	51
May 2025	24	5	44	73
Jun 2025	49	4	8	61
Jul 2025	23	5	1	29

Viewed (geographical distribution)

Total article views: 889 (including HTML, PDF, and XML) Thereof 889 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 24 Jul 2025

Download

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Preprint (4489 KB)
Metadata XML

Short summary

The carbon uptake period (CUP) refers to the time of the year when there is net absorption of CO₂ from the atmosphere to the land. Several studies have assessed changes in CUP based on seasonal metrics from CO₂ mole fraction measurements to understand the response of terrestrial biosphere to climate variations. However, we find that the CUP derived from CO₂ mole fraction measurements are not likely to provide an accurate magnitude of the actual changes occurring over the surface.


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

How atmospheric CO2 can inform us on annual and decadal shifts in the biospheric carbon uptake period

Journal article(s) based on this preprint

Interactive discussion

Interactive discussion

Peer review completion

Suggestions for revision or reasons for rejection

Suggestions for revision or reasons for rejection

Journal article(s) based on this preprint

Viewed

Viewed (geographical distribution)

How atmospheric CO₂ can inform us on annual and decadal shifts in the biospheric carbon uptake period