Representing high-latitude deep carbon in the pre-industrial state of the ORCHIDEE-MICT land surface model (r8704)

Xi, Yi; Ciais, Philippe; Zhu, Dan; Qiu, Chunjing; Zhang, Yuan; Peng, Shushi; Hugelius, Gustaf; Bowring, Simon P. K.; Goll, Daniel S.; Wang, Ying-Ping

doi:https://doi.org/10.5194/gmd-2024-206

Preprints

https://doi.org/10.5194/gmd-2024-206

Preprints

Submitted as: development and technical paper

23 Jan 2025

Submitted as: development and technical paper |

| 23 Jan 2025

Status: a revised version of this preprint is currently under review for the journal GMD.

Representing high-latitude deep carbon in the pre-industrial state of the ORCHIDEE-MICT land surface model (r8704)

Yi Xi, Philippe Ciais, Dan Zhu, Chunjing Qiu, Yuan Zhang, Shushi Peng, Gustaf Hugelius, Simon P. K. Bowring, Daniel S. Goll, and Ying-Ping Wang

Abstract. Field measurements, after extrapolation, suggest that deep Yedoma deposits (formed during the Pleistocene) and peatlands (formed during the Holocene) account for over 700 Pg C of soil carbon storage. Incorporating this old, deep, cold carbon into land surface models (LSMs) is crucial for accurately quantifying soil carbon responses to future warming. However, it remains underrepresented or absent in current LSMs, which typically include a passive soil carbon pool to represent all 'old carbon' and lack the vertical accumulation processes that deposited deep carbon in the soil layers of peat and Yedoma. In this study, we propose a new, more realistic protocol for simulating deep and cold carbon accumulation in the high latitudes, using the ORCHIDEE-MICT model. This is achieved by 1) integrating deep carbon from Yedoma deposits whose formation is calculated using Last Glacial Maximum climate conditions, and 2) prescribing the inception time and location of northern peatlands during the Holocene using spatially explicit data on peat age. Our results show an additional 157 Pg C in present-day Yedoma deposits, as well as a shallower peat carbon depth (by 1–5 m) and 35 Pg C (43 %) less passive soil carbon in northern peatlands, compared to the old protocol that ignored Yedoma deposits and applied a uniform, long-duration (13,500 years) peat carbon accumulation across all peatlands. As a result, the total organic carbon stock across the Northern Hemisphere (> 30° N) simulated by the new protocol is 2,028 Pg C, which is 226 Pg C higher than the previous estimate. Despite the significant challenge in simulating deep carbon with ORCHIDEE-MICT, the improvements in the representation of carbon accumulation from this study provide a model version to predict deep carbon evolution during the last glacial-deglacial transition and its response to future warming. The methodology implemented for deep carbon initialization in permafrost and cold regions in ORCHIDEE-MICT is also applicable to other LSMs.

Received: 05 Nov 2024 – Discussion started: 23 Jan 2025

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, 4226 KB)

Supplement (8623 KB)

Download & links

Yi Xi, Philippe Ciais, Dan Zhu, Chunjing Qiu, Yuan Zhang, Shushi Peng, Gustaf Hugelius, Simon P. K. Bowring, Daniel S. Goll, and Ying-Ping Wang

Status: final response (author comments only)

RC1:
'Comment on gmd-2024-206', Anonymous Referee #1, 17 Mar 2025
This paper introduces new model processes to represent old and deep carbon formation in Yedoma and peatland soils. It is an interesting model development, as it is a common issue for many land surface or ecosystem models to simulate these deep and old carbon dynamics explicitly. However, I struggled to understand the model comparisons, evaluation, and calibration due to the current structures and the scattered information. Here are some of my major points:
The authors merged different versions of ORCHIDEE models, so it seems unclear if any independent calibration or evaluation has been conducted before comparing carbon accumulations to ensure the model's performance is on the right track before adding new processes.

It is a bit confusing for me what has been developed in this study in terms of yedoma deposits. In the method section, it states that this study merged yedoma process from Zhu et al., 2016, and then in the Results section on L275, the authors used the same survey to evaluate the model performance. In Figure 4, why not compare the simulation outputs between the old and new spinup to show the model differences?

When comparing the spatial distribution of simulated carbon accumulation and maximum depth, why not compare the existing datasets to illustrate whether the changes make the simulations closer to the observation-based patterns? I later saw the comparison of the new spinup output with the existing database in Figure 9 in the Dicussion section, which makes it difficult to judge the performance of these new model developments.

General comments: the authors may want to consider merging the Results and Discussion sections, as some content in the Discussion were in the Results section and vice versa.
What happens in the model (regarding plant and soil carbon pools) when converting an upland cell fraction to a peatland fraction? It could be good to describe so readers can better understand the mechanism of these new changes in dynamic peatland areas and peatland inception time.
Some detailed comments:
Abstract:
L21: “A passive soil carbon pool” could be rephrased. For Non-specialists in this field, we won’t know what “passive” means here. It can be easily misunderstood as it is not biologically active or is not contributing to any fluxes. The same goes to L27..What does “less passive” really mean?
Figure 1, &L148, what are peat PFT and Yedoma PFT? What are the main differences between these two PFTs? So why only one PFT is allowed to grow on peatland or Yedoma? How PFT chose could influence carbon accumulation over this study's long temporal scales?
L276, what is the value of total SOC stock from the old spin-up then?
L306-307, what are the main reasons that two spinup end with different NEP values?
Citation: https://doi.org/10.5194/gmd-2024-206-RC1
- AC1: 'Reply on RC1', Yi Xi, 09 May 2025
  
  Dear Reviewer,
  Thank you very much for taking the time to review our manuscript and for your thoughtful and constructive comments. Please find attached our detailed response to your suggestions.
  Best regards,
  
  Yi
  
  Citation: https://doi.org/10.5194/gmd-2024-206-AC1
CEC1:
'Comment on gmd-2024-206 - No compliance with the policy of the journal', Juan Antonio Añel, 24 Mar 2025

Dear authors,
Unfortunately, after checking your manuscript, it has come to our attention that it does not comply with our "Code and Data Policy".

https://www.geoscientific-model-development.net/policies/code_and_data_policy.html
You have archived the ORCHIDEE code in a site that does not comply with our requirements. Therefore, you must store the code in a repository that complies with our policy, and provide a link and permanent identifier (e.g. DOI or Handle).
A similar issue happens with the WISE30sec and the NCSDC data. It would be good if you could contact their authors/maintainers and request them to store the data in an acceptable repository, and then reply to this comment with the information.
Please, reply to this comment as soon as possible with the information about it, and modify the Code Availability section of your manuscript accordingly in future versions of your manuscript.
Juan A. Añel

Citation: https://doi.org/10.5194/gmd-2024-206-CEC1
- CC1:
  'Reply on CEC1', Daniel S. Goll, 24 Mar 2025
  
  Dear Juan
  thanks for pointing this issue out. We weren't aware of the shift in the execution of GMD policy. We will do as proposed on the GMD website
  Best wishes
  Daniel
  
  Citation: https://doi.org/10.5194/gmd-2024-206-CC1
  - AC3: 'Reply on CC1', Yi Xi, 10 May 2025
    
    Dear Juan,
    
    Thank you for pointing out this important issue.
    
    We have archived the model code in a public repository on Zenodo (https://zenodo.org/records/15306029) and cited it in the revised manuscript to ensure reproducibility.
    
    Regarding the WISE30sec and NCSCD datasets, we did not request the original authors to deposit them in an alternative repository, as these datasets have been widely used over the past decade and are consistently cited using the established links. Changing those links might disrupt continuity for previous studies. Instead, to support transparency and reproducibility, we have uploaded the subset of these datasets used in our study (limited to our study domain and spatial resolution), along with all processing scripts and code used to generate the results and figures in our study, to a separate public Zenodo repository (https://zenodo.org/records/15371113). This has also been cited in the revised manuscript.
    
    If this approach does not meet the data availability standards, we will try to reach out to the original dataset authors for any possibility of archiving the data in a compliant repository, though we cannot guarantee a positive answer.
    
    We hope this solution is acceptable, and we appreciate your understanding.
    
    Best regards,
    Yi
    
    Citation: https://doi.org/10.5194/gmd-2024-206-AC3
RC2:
'Comment on gmd-2024-206', Anonymous Referee #2, 26 Mar 2025
This paper summarizes model development which I think tackles important points: Yedoma C stocks and peatland C stocks. I think the model structure is fine, given that the earlier versions are published. This is why I give the manuscript low marks for scientific significant. I am also grading it lowly for scientific quality for the reasons I outline below. Scientific reproducibility: it's been noted not to be in compliance with data policy. Presentation quality is okay.
Major concerns:
This paper seems to represent an important and perhaps technically challanging but scientifically, somewhat incremental advancement in the model, having been the result of harmonizing several different model branches.

There is a spatial mis-match between the two domains (Yedoma and peatlands) which is confusing and never explicitly stated until Table 3. The peatland model simulation includes not only northern temperate, boreal, and Arctic peatlands but also southern temperate and sub-tropical peatlands and the simulation extends to 30 N. The Hugelius et al. (2020) peatland extent dataset does not go all the way to 30 N. These southern peatlands are also simulated in response to de-glaciation, but this is not a driving factor in peatland initiation for these peatlands at the southern margin of the temperate zone (Treat et al., 2019). A map of grid-cell level peatland initiation would help to clarify this approach. With the extension of modeling to more southern regions, this starts to get challenging because of anthropogenic peatland drainage during the last millennium (e.g. Fluet-Chouinard et al., 2022) which points out the vast extent of wetlands and peatlands lost to drainage. However, this is not accounted for in the past wetland C extent when it is based on present-day areas from Hugelius. This is very problematic.

This is a publication is submitted now, in 2025, but most of the references in the introduction are out of date and, even more importantly, the datasets used in the analysis and as model inputs are also out of date. These need to be updated to include the most recent research. This would be desirable in any case but I think this is really imperative given major point #1. Consider, for example, the dataset presented in Treat et al. (2019) contains over 3900 basal ages of peatland initiation plus information on C accumulation in peatlands over time and a comparison between spatial extent and C stock. Hugelius et al. (2020) contains new information on peatland C stocks. Mishra et al. (2021) has additional analysis of permafrost C stocks. Strauss et al. (2017) and Strauss et al. (2021) have updated information about Yedoma stocks. Treat et al. (2021) has information on spatial coverage and evolution of northern peatlands over time. Brosius et al. (2021) include lake initiation ages for Yedoma thermokarst lakes, which is in some confusing way alluded to in the description of peatlands. Kleinen et al. (2012) and 2016 also describe peatland initiation and C storage development during the Holocene, which has also been done empirically by Nichols & Peteet (2019). The list goes on, these are only suggestions, but an improved literature review is needed.

Abstract
Define deep yedoma deposits

Define spatial extent of analysis

What about CLM and earlier Orchidee where this is done for permafrost?

Is C stocks all that matter?

Intro: some of the numbers and references for deep C stocks in the north seem out of date (e.g. there are newer papers). Same with Yedoma. Same with tropical peatlands. Many of the cited references are over 10 years old, there are many new developments and accounting that has happened since! The only more recent one is Hugelius et al 2020!

Methods peatlands
Basal ages: Treat et al. 2019/2021 seems more appropriate with a focus on permafrost ecosystems.
Addition of info on yedoma here is confusing. Why is the yedoma in the peatlands session? Why don’t you use the Brosius et al dataset that looks at lake formation? Oh, it’s just to justify peatlands, that yedoma turns off and peatlands turn on?
Methods Yedoma/peatland is also a bit tricky, there is this Holocene cover on top of Yedoma deposits that can be 1-2m thick that I can’t quite figure out how is dealt with in this modeling approach. This is much more C rich than Yedoma.

Results
Yedoma, what about Jens synthesis?
Super high C densities, is this really considering ice density and ice and ice wedges? Seems higher than estimates that I’ve seen from Hugelius et al.which could be an average.
Peat PFT is a bit high, check temporal trends, seems too little in the initial stages.

Difference in spatial domains is really confusing. Only from Table 3 does it become clear that the peat simulation extends to 30 N but the Ydoma simulation is clearly not relevant here.

References
Brosius, L.S., K.M. Walter Anthony, C.C. Treat, J. Lenz, M.C. Jones, M.S. Bret-Harte, G. Grosse (2021). Spatiotemporal patterns of northern lake formation since the Last Glacial Maximum. Quaternary Science Reviews 253, 106773. doi: 10.1016/j.quascirev.2020.106773
Hugelius, G., J. Loisel, S. Chadburn, R. B. Jackson, M. Jones, G. MacDonald, M. Marushchak, D. Olefeldt, M. Packalen, M. B. Siewert, C. Treat, M. Turetsky, C. Voigt and Z. Yu (2020). Large stocks of peatland carbon and nitrogen are vulnerable to permafrost thaw. Proceedings of the National Academy of Sciences: 201916387. doi: 10.1073/pnas.1916387117
Mishra, U., G. Hugelius, E. Shelef, Y. Yang, J. Strauss, A. Lupachev, J. Harden, J. Jastrow, C.-L. Ping, W. Riley, E.A.G. Schuur, R. Matamala, M. Siewert, L. Nave, C. Koven, M. Fuchs, J. Palmtag, P. Kuhry, C. Treat, S. Zubrzycki, F. Hoffman, B. Elberling, P. Camill, A. Veremeeva, A. Orr (2021). Spatial heterogeneity and environmental predictors of permafrost region soil organic carbon stocks. Science Advances 7 (9): eaaz5236. doi: 10.1126/sciadv.aaz5236.
Kleinen, T., Brovkin, V., and Schuldt, R. J.: A dynamic model of wetland extent and peat accumulation: results for the Holocene, Biogeosciences, 9, 235–248, https://doi.org/10.5194/bg-9-235-2012, 2012.
Kleinen, T., Brovkin, V., and Munhoven, G.: Modelled interglacial carbon cycle dynamics during the Holocene, the Eemian and Marine Isotope Stage (MIS) 11, Clim. Past, 12, 2145–2160, https://doi.org/10.5194/cp-12-2145-2016, 2016.
Treat, C. C., T. Kleinen, N. Broothaerts, A. S. Dalton, R. Dommain, T. A. Douglas, J. Z. Drexler, S. A. Finkelstein, G. Grosse, G. Hope, J. Hutchings, M. C. Jones, P. Kuhry, T. Lacourse, O. Lähteenoja, J. Loisel, B. Notebaert, R. J. Payne, D. M. Peteet, A. B. K. Sannel, J. M. Stelling, J. Strauss, G. T. Swindles, J. Talbot, C. Tarnocai, G. Verstraeten, C. J. Williams, Z. Xia, Z. Yu, M. Väliranta, M. Hättestrand, H. Alexanderson and V. Brovkin (2019). "Widespread global peatland establishment and persistence over the last 130,000 y." Proceedings of the National Academy of Sciences: 201813305.
Treat, C.C., Jones, M.C., Brosius, L., Grosse, G., Anthony, K.W. and Frolking, S., 2021. The role of wetland expansion and successional processes in methane emissions from northern wetlands during the Holocene. Quaternary Science Reviews, 257, p.106864.
Citation: https://doi.org/10.5194/gmd-2024-206-RC2
- AC2: 'Reply on RC2', Yi Xi, 09 May 2025
  
  Dear Reviewer,
  Thank you very much for taking the time to review our manuscript and for your thoughtful and constructive comments. Please find attached our detailed response to your suggestions.
  Best regards,
  
  Yi
  
  Citation: https://doi.org/10.5194/gmd-2024-206-AC2
RC3:
'Comment on gmd-2024-206', Anonymous Referee #3, 05 May 2025

This paper introduces a major advancement in representing deep carbon storage in high-latitude soils, specifically in Yedoma deposits and peatlands, within the ORCHIDEE-MICT land surface model. Deep soil organic carbon (SOC) plays a crucial role in understanding carbon cycle responses to climate change, yet it has been inadequately represented in many existing land surface models. The authors address this gap by developing a new simulation protocol that incorporates realistic sedimentation processes for Yedoma (originating during the Pleistocene) and a time-explicit approach to peat accumulation (formed during the Holocene).
The improved model estimates an additional 157 Pg C stored in Yedoma and refines peatland carbon stocks, resulting in a total increase of 226 Pg C in simulated SOC across the Northern Hemisphere (>30°N) compared to previous versions. This enhanced representation captures both the spatial and vertical complexities of SOC distribution more accurately, providing a stronger foundation for predicting future carbon dynamics in permafrost regions.
A key innovation is the peatland soil tile, introduced by Qiu et al. (2018; 2019), which specifically models peatland environments characterized by saturated, oxygen-deficient, and nutrient-limited conditions. This tile allows for improved simulation of carbon accumulation under waterlogged conditions, which greatly slows decomposition and promotes deep organic matter storage.
The new model setup departs from traditional, static assumptions used in conventional land surface model spin-ups. Instead, it integrates process-based, historically grounded reconstructions of deep carbon formation, resulting in a more realistic initialization of high-latitude carbon stocks.
Initialization and spin-up procedures are often underappreciated aspects of land surface modeling, yet they are critical for improving the accuracy of permafrost carbon feedback projections. This study represents an important and timely contribution to the field, highlighting the complexity of SOC modeling in high-latitude environments. It is well-written and comprehensive, and I have no further comments. I recommend this manuscript for immediate publication.

Citation: https://doi.org/10.5194/gmd-2024-206-RC3
- AC4: 'Reply on RC3', Yi Xi, 10 May 2025
  
  Dear Reviewer,
  
  Thank you very much for taking the time to review our manuscript and for your encouraging comments. We greatly appreciate your recognition of the significance and timeliness of our work.
  
  Following the suggestions from the other two reviewers, we have further refined the manuscript by clarifying our model development and evaluation strategies. We have also expanded the model description to include more details on changes in plant and soil carbon pools associated with the conversion from conventional soils to peatlands. In addition, we have added further discussion on the limitations of our approach.
  
  We believe these revisions have significantly improved the clarity and comprehensiveness of the manuscript. Thank you again for your thoughtful review!
  
  Best regards,
  Yi
  
  Citation: https://doi.org/10.5194/gmd-2024-206-AC4

Yi Xi, Philippe Ciais, Dan Zhu, Chunjing Qiu, Yuan Zhang, Shushi Peng, Gustaf Hugelius, Simon P. K. Bowring, Daniel S. Goll, and Ying-Ping Wang

Supplement

https://doi.org/10.5194/gmd-2024-206-supplement

Yi Xi, Philippe Ciais, Dan Zhu, Chunjing Qiu, Yuan Zhang, Shushi Peng, Gustaf Hugelius, Simon P. K. Bowring, Daniel S. Goll, and Ying-Ping Wang

Viewed

Total article views: 366 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
279	68	19	366	40	13	11

HTML: 279
PDF: 68
XML: 19
Total: 366
Supplement: 40
BibTeX: 13
EndNote: 11

Views and downloads (calculated since 23 Jan 2025)

Month	HTML	PDF	XML	Total
Jan 2025	60	14	5	79
Feb 2025	76	17	5	98
Mar 2025	74	15	3	92
Apr 2025	26	9	1	36
May 2025	43	13	5	61

Cumulative views and downloads (calculated since 23 Jan 2025)

Month	HTML	PDF	XML	Total
Jan 2025	60	14	5	79
Feb 2025	76	17	5	98
Mar 2025	74	15	3	92
Apr 2025	26	9	1	36
May 2025	43	13	5	61

Viewed (geographical distribution)

Total article views: 365 (including HTML, PDF, and XML) Thereof 365 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 15 May 2025

Short summary

Including high-latitude deep carbon is critical for projecting future soil carbon emissions, yet it’s absent in most land surface models. Here we propose a new carbon accumulation protocol by integrating deep carbon from Yedoma deposits and representing the observed history of peat carbon formation in ORCHIDEE-MICT. Our results show an additional 157 PgC in present-day Yedoma deposits and a 1–5 m shallower peat depth, 43 % less passive soil carbon in peatlands against the convention protocol.


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