Modeling soil organic carbon dynamics in temperate forests with Yasso07

Mao, Zhun; Derrien, Delphine; Didion, Markus; Liski, Jari; Eglin, Thomas; Nicolas, Manuel; Jonard, Mathieu; Saint-André, Laurent

doi:https://doi.org/10.5194/bg-16-1955-2019

Articles | Volume 16, issue 9

https://doi.org/10.5194/bg-16-1955-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/bg-16-1955-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 16, issue 9

Research article

|

13 May 2019

Research article |

| 13 May 2019

Modeling soil organic carbon dynamics in temperate forests with Yasso07

Zhun Mao, Delphine Derrien, Markus Didion, Jari Liski, Thomas Eglin, Manuel Nicolas, Mathieu Jonard, and Laurent Saint-André

Download

Final revised paper (published on 13 May 2019)
Supplement to the final revised paper
Preprint (discussion started on 18 Jun 2018)
Supplement to the preprint

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Review of Mao et al', Anonymous Referee #2, 06 Jul 2018
- AC1: 'responese to comments', Zhun Mao, 31 Aug 2018
RC2: 'Problem with scaling quality of steady state stocks', Thomas Wutzler, 30 Jul 2018
- AC2: 'responese to comments', Zhun Mao, 31 Aug 2018

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

ED: Publish subject to minor revisions (review by editor) (11 Sep 2018) by Luo Yu

ED: Reconsider after major revisions (09 Oct 2018) by Luo Yu

AR by Svenja Lange on behalf of the Authors (12 Oct 2018) Author's response

ED: Referee Nomination & Report Request started (12 Oct 2018) by Luo Yu

RR by Thomas Wutzler (15 Oct 2018)

Suggestions for revision or reasons for rejection

The revised paper acknowledges my main concern of improper initialization of carbon quality distribution by a short discussion. The authors argue that comparing another initialization method would be too much work and should go to a new paper.
However, I still think that the entire part on annual carbon accumulation (ACC) cannot be trusted with the currently applied assumption of initializing carbon quality distribution. This might be only my personal interpretation that I defend below, but the authors should better defend their assumption and the validity of their results. Or they should clearly acknowledge that only an upcoming paper will supply more reliably conclusions. For example, one of the conclusions is that its “Yasso07s failure of too much penalizing loss of labile carbon” (P21L4ff). I disagree. To my view it’s the author’s assumptions of scaling the more labile carbon stocks during initialization that determines the ACC pattern in the simulations and its missing correlation with observations.

Thanks for providing Fig S4. From lines p14/22ff of the manuscript, I infer that „simulated“ corresponds to the equilibrium carbon stocks and „observed“ corresponds to measured 40cm stocks at initial time extrapolated down to 1m, correct? This should be clarified in the figure title. Moreover, mentioning ACC in the figure title is probably an error here.
The figure shows that for most broadleaved, observed stocks are lower than equilibrium, which could indicate that equilibrium is not reached after last disturbance yet.
For the conifers it shows that observed stocks are larger than simulated equilibrium, hinting to some model inadequacy, or mismatches in input fluxes, or underestimate of stead state stocks, e.g. by a too high decomposition rate of the slow pool.

From comparing Fig S4 and Fig. 3 I see that where initial stocks are overestimated, also the stock change is overestimated, and similarly where initial stocks were underestimated, stock change is underestimated. To me, this hints to the suspected large effect of scaling faster pools when transferring steady state to observed stocks.

Fig. 4 could probably also be explained by the same issue. When initial steady state stocks are higher than observed stocks, you downscale the faster pools, leading to stronger positive decadal ACC when pools develop towards steady state again. Similarly, when initial steady state stocks are lower, you upscale the faster pools leading to negative decadal ACC during simulation. The missing of this pattern in observations suggests to me to better not scale the faster pools.

My conclusion from you results, therefore, is that probably the mismatch in initial distribution of qualities determines your pattern of carbon accumulation. To me most of the results come down to the assumption of keeping carbon quality distribution constant when adjusting stocks.

Specific comments:

I cannot get the message from Fig. S9. However, it should be crucial to my interpretation of importance of changes to steady state litter quality distribution for ACC. Can you simplify it?

P22l15ff: Thanks for including discussion on the relaxed equilibrium assumption. Lack of information on modified rate of H pool should not be a problem at the time scale of interest. Just use a significantly slower decomposition rate, as suggested by Wutzler 2007. The difference will only be relevant at longer time scale.

P23L7: Note that the intended usage of depth-dependent decomposition rates renders the calibrated model only applicable at the site of calibration. Depth is often just a surrogate factor for other stabilization mechanisms. Trying to capture better indicators for these stabilization mechanisms is better than relying on their relation with depth that will change across sites.

Hide

RR by Anonymous Referee #2 (22 Oct 2018)

ED: Reconsider after major revisions (27 Nov 2018) by Luo Yu

AR by Anna Wenzel on behalf of the Authors (29 Jan 2019) Author's response Manuscript

ED: Referee Nomination & Report Request started (13 Feb 2019) by Luo Yu

RR by Thomas Wutzler (18 Feb 2019)

RR by Anonymous Referee #2 (28 Mar 2019)

ED: Publish subject to technical corrections (01 Apr 2019) by Luo Yu

AR by Zhun Mao on behalf of the Authors (09 Apr 2019) Author's response Manuscript

Post-review adjustments

AA: Author's adjustment | EA: Editor approval

AA by Zhun Mao on behalf of the Authors (09 May 2019) Author's adjustment Manuscript

EA: Adjustments approved (09 May 2019) by Luo Yu

Download

Article (4788 KB)
Full-text XML

Short summary

In a context of global changes, modeling and predicting the dynamics of soil carbon stocks in forest ecosystems are vital but challenging. Yasso07 is considered to be one of the most promising models for such a purpose. We examine the accuracy of its prediction of soil carbon dynamics over the whole French metropolitan territory at a decennial timescale. We revealed how the bottleneck in soil carbon modeling is linked with the lack of knowledge on soil carbon quality and fine-root litter.