Assessing the response of soil carbon in Australia to changing inputs and climate using a consistent modelling framework

Lee, Juhwan; Viscarra Rossel, Raphael A.; Zhang, Mingxi; Luo, Zhongkui; Wang, Ying-Ping

doi:https://doi.org/10.5194/bg-18-5185-2021

Articles | Volume 18, issue 18

https://doi.org/10.5194/bg-18-5185-2021

© Author(s) 2021. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/bg-18-5185-2021

© Author(s) 2021. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 18, issue 18

Research article

|

22 Sep 2021

Research article |

| 22 Sep 2021

Assessing the response of soil carbon in Australia to changing inputs and climate using a consistent modelling framework

Juhwan Lee, Raphael A. Viscarra Rossel, Mingxi Zhang, Zhongkui Luo, and Ying-Ping Wang

Download

Final revised paper (published on 22 Sep 2021)
Supplement to the final revised paper
Preprint (discussion started on 21 Jul 2020)
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 comment', Anonymous Referee #2, 31 Jul 2020
- AC1: 'Response to referee#2', Raphael Viscarra Rossel, 25 Sep 2020
RC2: 'Report for bg-2020-150', Anonymous Referee #1, 07 Aug 2020
- AC2: 'Response to referee #1', Raphael Viscarra Rossel, 25 Sep 2020

Peer-review completion

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

ED: Reject (27 Sep 2020) by Luo Yu

ED: Reconsider after major revisions (23 Oct 2020) by Fortunat Joos

Dear authors

The handling editor decided to reject your paper following the recommendations of two referees. This decision was overturned by the Chief Editors following your appeal to the editorial board. I was asked by the editorial board to handle your manuscript.

I am happy to invite you to submit a suitably revised version of your manuscript, considering the comments by the two referees and my comments. Your revised manuscript will be sent out again to review.

Please carefully consider the comments made by the two reviewers and change your manuscript to address these points where appropriate.

I suggest that you add a few lines early on in the manuscript to clarify the setup of your sensitivity simulations with altered carbon input to soils under constant climate and provide an explanation/justification on the range of input fluxes selected to address the second and third main point of referee #2. In this context, please consider replacing the word “prediction/predict” with “simulations/simulate” or similar to avoid the impression that results represent a prediction of the future evolution of soil C in Australia.

Reviewer #1 perceived your work and presentation as very technical. You kindly offered to further emphasize the science questions in your reply to reviewer #1. It would, in my opinion, indeed be useful to provide a few additional lines further describing the context and motivation for your model framework and your sensitivity simulations in the introduction, thereby complementing the context and motivation already provided. I may be wrong, but I got the impression that one motivation of this study is to prepare the ground for further application of your model (i) in assessments of Australian National Greenhouse Gas Inventory (line 50) used for the reporting under the UNFCCC process (?) and (ii) to pay farmers for efforts to stimulate extra carbon transfer to soils (line 51). I also interpret your text on line 51, that your model may be used to estimate the change in soil carbon stock over a 100-yr period under constant climate in response to measures by farmers, and in turn, this estimate is used to pay farmers. In any case, I would welcome further motivation in the manuscript why you are performing this model exercise, e.g., also building on your answer to reviewer #1 on page C5 in the reply (“Our work addresses important scientific questions pertinent to Australian soils: e.g. how…”).

Reviewer #1 noted that section 2.3.4 is unclear. The description of your iterative procedure to estimate soil C input in section 2.3.4 should indeed be improved. What do you mean exactly by “We tested six different values of DPM/RMP ratio … We then perform the simulation iteratively …” Did you run for each value of DPM/RPM an iteration to estimate C input for a given DPM/RMP value? The sentence on l. 169/170 is unclear: “We ran the model for 100 years [ ] until equilibrium conditions occurred.” Maybe you want to say that you run the model several times for 100 yr with slightly different C input until equilibrium is achieved? Does this also imply that the TOC inventory at the end of the iterative procedure is the same as prescribed at the beginning of the iteration (apart from the bacterial pool? Where the soil pools initialized again with the measurements after each simulation in the iteration? Please describe your simulation protocol more precisely.

Regarding the equilibrium assumption, it is not clear to me why you need to estimate soil input iteratively. Could this not be done by setting C input equal to all C loss fluxes?

In general, I found your proposed text modifications appropriate in response to the comments of referee #2. Please implement these changes when revising your manuscript for resubmission. Please also consider whether additional clarifications may help to avoid misunderstanding and address comments where you did not indicate any action in your response

I am looking forward to receiving your revised manuscript. Thank you for submitting your work to Biogeosciences.

Yours sincerely,

Fortunat Joos

Further minor comments:

L80: Please provide the reference state (Temp, soil water veg. cover, ..) for the specified values of the decomposition rate coefficients.

261/262: please clarify this sentence. It is unclear why an increase of 39% is larger than an increase of 59%?

Figure 3: I had the same issue with overlapping labels as referee #2. Please check your figure and, perhaps, select slightly smaller fonts.

Hide

AR by Raphael Viscarra Rossel on behalf of the Authors (11 Dec 2020) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (16 Dec 2020) by Fortunat Joos

RR by Iain Colin Prentice (13 Jan 2021)

RR by Anonymous Referee #4 (28 Jan 2021)

Suggestions for revision or reasons for rejection

Review of MS bg-2020-150 “Simulation of soil carbon dynamics in Australia with Rᴏᴛʜ C” by Lee et al.

I reviewed the revised version of the MS, also taking note of the previous reviews and changes that were made to the new version.

This is a very interesting, technically demanding study where the controlling parameters input rate and input quality for soil carbon stocks in Australian ecosystems were evaluated by using RothC and an elaborated set of external data. The overall structure of the work, its general quality, and the fact that it seamlessly follows a number of previous Australian studies on the same topic, where different data sets and methods have been developed that were also used here, is appreciated.

In line with a previous comment, I see a major drawback in that the authors did not use climate change (CC) scenarios for calculating the effect of increasing C inputs to SOC stocks. Considering the well documented effect that CC will have on future SOC in previous modeling approaches from other regions, it seems not plausible to rely on a repetition of time windows of past climate to simulate the response of SOC to changing conditions which, by nature, can only take place in the future. I therefore suggest that the authors use CC scenarios for evaluating the effect of higher C input on SOC storage. In this context, also the discussion starting in line 362 could be extended towards climate as an important determinant of long-term SOC changes.

Estimating or simulating belowground input is crucial for every SOC modeling study. On pages 6-7, the approach used for estimating these inputs is described. A fixed shoot:root ratio is considered to get the belowground part from the crop model (whether or not shoot:root ratios vary for a specific crop depending on external conditions such as management is a matter of debate in the literature [e.g. Hirte et al. 2018], but such a simplified approach seems justified given the poor availability of belowground input data from experiments). The authors then change monthly inputs (page 8) to better understand the response of the SOC pool to it. In consequence, shoot:root ratios will change as well. I suggest to display and discuss the resulting shoot:root ratios, as they provide a means of quality control to the overall approach. Reviews such as the one from Bolinder et al. (2007) might be helpful to put the derived ratios into context.

The selected range of inputs of between 0.25 – 6 times the equilibrium input seems wide, and Fig. 4 indicates inputs of up to 10 t C ha-1 per year. Even though the authors acknowledge that allocating inputs to sites/soils where higher additional storage might be achieved (line 360) is aimed for, an increase by more than 2-fold seems highly unrealistic given issues of transport or nutrient input. For comparison, a recent study from the temperate zone came up with an estimate of c. 3.7 t C ha-1 for croplands and grasslands (Jacobs et al. 2020). Therefore, I consider these input scenarios as partially unrealistic and suggest to downscale them a bit.

References
Bolinder, M.A., Janzen, H.H., Gregorich, E.G., Angers, D.A., Vandenbygaart, A.J., 2007. An approach for estimating net primary productivity and annual carbon inputs to soil for common agricultural crops in Canada. Agriculture Ecosystems & Environment 118, 29-42.
Hirte, J., Leifeld, J., Abiven, S., Oberholzer, H.-R., Mayer, J., 2018. Below ground carbon inputs to soil via root biomass and rhizodeposition of field-grown maize and wheat at harvest are independent of net primary productivity. Agriculture, Ecosystems & Environment 265, 556-566.
Jacobs, A., Poeplau, C., Weiser, C., Fahrion-Nitschke, A., Don, A., 2020. Exports and inputs of organic carbon on agricultural soils in Germany. Nutrient Cycling in Agroecosystems 118, 249-271.

Hide

RR by Anonymous Referee #5 (09 Feb 2021)

Suggestions for revision or reasons for rejection

Abstract: The opening of the abstract is hard to follow, I can imagine any number of “disconnects” not disconnections, but the abstract doesn’t state the problem being addressed. I also think this is a technical issue and put a modeling issue before the science question. The abstract should begin with a clear statement about what the authors are actually studying. Something like “Soil organic matter is a key reservoir for carbon, and its vulnerability to global change depends also om many anthropogenic management activities”..

This is pretty generic:
“However, a better understanding of soil organic C dynamics is needed to determine the size of the soil C pool accurately and to assess the potential for those opportunities”.
Can you be more specific about why better understanding is needed? What are the big uncertainties? This statement can be said about anything, but without specifics, why bother saying this? This is an opportunity to be specific, not generic.

Is science a popularity contest? Is this really what you mean to say?
Other reasons for their continued popularity might be that there is ample documentation on them; they are relatively simple and general and are therefore also well understood.
I would think, and it would certainly strengthen this paper if Roth-C and Century were not popular because they are simple and easy to use but because they are well-tested, accurate and robust (which I believe they are), and I believe they’re little evidence that more complex models are more accurate for this type of study.

A key part of this paper is the integration of the ROTH-C model and data, yet the introduction is strangely vague on this subject. What deficiencies arise from poor data synergy, and what approach is proposed? The paragraph around Line 60 is very superficial.

Figure 1: How is this different? The input data sets look very standard and can the figure show what’s novel in this approach? Eg: remote sensing, spectroscopy, etc

I always discourage statements like the following:
This gave us confidence in the performance of the model in Australia.
What does “confidence” mean? Good enough for a paper, good enough to make management recommendations with financial conbsequences? I much preferred statements about the likely uncertainty, and openness about where and when the model could fail! Confidence is like belief, it has very limited place in a scientific paper!

Figure 3: these model fits are SO good they raise concern, when simulating so notoriously variable a quantity as SOC, some discussion of why and whether that would be expected away from the calibration sites! In the caption, it states after optimization but it would be very easy for a casual reader to assume this is the likely model accuracy, especially after the “confidence statement”.

Figure 4 and 5 are very nice. The estimate of importance in Figure 6 is quite unclear, and could be expanded or even defined in the caption.

Around Line 305: I suggest avoiding “performed well” “additional confidence” and replace all of these statements that are subjective with explicit statements of model skill. Statements about “performing well” or “confidence” should be reserved for the very end of the discussion where the use of the model and output for managers is mentioned. In the results and discussion, focus on skill metrics, not subjective value statement.

This statement remains difficult for me. What is the disconnect? By this point statements like “Soil measurements are very local and sparse, yet soil information is required at farm to national levels”. Not “disconnect”. There are specific challenges, not well-sumarized by “disconnect”. This whole paragraph starting section 4.2 is redundant with earlier statements and adds no detail. What do these data sets add? I can see the authors grasping for a point here, but this entire section could be summarized by “Incorporating both local and mapped information enables the model to be run for large areas, or with more specificity for particular sites”. This has been done many times with ROTH-C (and Century)—these authors have developed a particular coding and data management approach, but it is very similar to gridded Century with its schedule files.

Future needs is so generic as to be useless

The conclusions are repetitious with early text and could be strengthened significantly.
There is a functional disconnection between measurements, data and biogeochemical models (Blankinship et al., 2018), but by simulating under a framework, like we did here, we can bridge that disconnect.

Hide

ED: Reconsider after major revisions (10 Feb 2021) by Fortunat Joos

Dear authors

Your manuscript was meanwhile assessed by three new reviewers. All three reviewers note that your work is of high technical quality and of importance but request improvements before a potential publication. The three reviewers offer points of criticisms and very valuable advice on how to further improve the manuscript. I ask you to carefully consider the issues raised by the reviewers and to prepare a revised manuscript and a point-by-point reply to the review comments. The manuscript will then be sent to the two reviewers that asked for major revisions.

Reviewer #3 asks that the underpinning data and codes are made publicly available. It is becoming more and more praxis and also requested by Biogeosciences to publish input data sets, results, and codes on a public server, preferentially with a doi attached. I ask you to follow this practice to the extent this is reasonable.

Reviewer #4 raises three important scientific issues.
(i) Estimating the effect of changing inputs on SOC stocks over a 100-year time horizon without consideration of climate change appears not meaningful. This is certainly a valid criticism as global warming is continuing. A relatively small effort, with results for a small number of climate scenarios and a 'mean' input, might inform the reader about the sensitivity of your results to different climate conditions. These additional results may be presented in an appendix and briefly discussed in the main text or presented in the main text.
(ii) The reviewer also recommends to discuss shoot-root ratios under altered input and to compare these ratios to literature values as an overall quality control.
(iii) Finally, reviewer #3 notes that some of the input scenarios are partially unrealistic. This point should at least be reflected in the discussion.

Reviewer #5 appreciates the technical aspects of your work but calls for an improved presentation. The reviewer critiques vague and subjective statements, and the lack of a clear definition of the problem, the outcomes, and the novelty. Please consider these comments very carefully and revise your manuscript accordingly.

Please follow the guidelines of the journal when preparing your response: The author's response in case of "minor" or "major" revisions must be submitted as one separate *.pdf file (indicating page and line numbers), structured in a clear and easy-to-follow sequence: (1) comments from referees/public, (2) author's response, and (3) author's changes in manuscript. Regarding author's changes, a marked-up manuscript version (track changes in Word, latexdiff in LaTeX) converted into *.pdf and combined with the author's response should be provided.

I am very much looking forward to receiving your revised manuscript. Thank you for submitting your work to Biogeosciences.

Yours sincerely,

Fortunat Joos

Hide

AR by Raphael Viscarra Rossel on behalf of the Authors (02 Aug 2021) Author's response Author's tracked changes

ED: Referee Nomination & Report Request started (04 Aug 2021) by Fortunat Joos

RR by D. Schimel (22 Aug 2021)

ED: Publish as is (23 Aug 2021) by Fortunat Joos

AR by Raphael Viscarra Rossel on behalf of the Authors (30 Aug 2021) Manuscript

Download

Article (4078 KB)
Full-text XML

Short summary

We performed Roth C simulations across Australia and assessed the response of soil carbon to changing inputs and future climate change using a consistent modelling framework. Site-specific initialisation of the C pools with measurements of the C fractions is essential for accurate simulations of soil organic C stocks and composition at a large scale. With further warming, Australian soils will become more vulnerable to C loss: natural environments > native grazing > cropping > modified grazing.