Physical protection of soil carbon stocks under regenerative agriculture

Keenor, Sam G.; Lee, Rebekah; Reid, Brian J.

doi:10.5194/soil-11-957-2025

Articles | Volume 11, issue 2

https://doi.org/10.5194/soil-11-957-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/soil-11-957-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 11, issue 2

Original research article

|

19 Nov 2025

Original research article |

| 19 Nov 2025

Physical protection of soil carbon stocks under regenerative agriculture

Sam G. Keenor, Rebekah Lee, and Brian J. Reid

Download

Final revised paper (published on 19 Nov 2025)
Supplement to the final revised paper
Preprint (discussion started on 22 Jan 2025)
Supplement to the preprint

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2024-4029', Anonymous Referee #1, 02 Mar 2025

This manuscript aims to determine the effects of regenerative agriculture on soil carbon stocks. The authors found that regenerative agriculture did not increase recalcitrant carbon stocks but did significantly increase labile carbon stocks. A lot of work has been done, and some results are interesting. However, I do not think the current version is suitable for publication in SOIL because of several weaknesses. Please see below for my main concerns:

First, the methods were not introduced clearly in the current manuscript, making it difficult to judge the results. For example, what were the management practices (fertilizer, tillage, and so on) that were used for the conventional farming? What were the grazing cover crops that were planted on the alleyways? What were the amount and nutrient content of the sprays of compost tea and organic fertilizer? Which year was the soil sample collected? There were many different soil carbon types in this manuscript (e.g. labile and recalcitrant carbon, occluded carbon, stabilised carbon, WSA, and NWSA); however, the authors failed to separate them clearly. It is hard for me to digest so many different terminologies.

Second, the main text should be improved seriously. For example, many parts in the Results and Discussion section were not about “results and discussion” but were about “background”, which should be moved to the introduction section. In addition, more information about the field design and methods should be included in the abstract; otherwise, it is difficult for readers to know the meaning of “alley soil and bush soil”.

Third, in the abstract section, the authors stated that “This research provides valuable insights into the mechanisms of soil carbon stabilisation under regenerative agriculture practices.” However, I am not sure if they really unravel the mechanisms because they did not measure other soil variables (e.g. soil nutrient content and soil microbial parameters), which can be used to explain the results. Moreover, the “Results and Discussion” section was mainly introducing background and describing the results. Instead, more deep discussions and measurements should be included to explain the interesting results (e.g. when compared to control treatment, why did total carbon stock decreased in the short term and then increased to the similar level like the control treatment in the long term). Does this mean regenerative agriculture must be conducted for a long time? Otherwise, croplands would lose soil carbon?

Citation: https://doi.org/10.5194/egusphere-2024-4029-RC1
- AC1:
  'Reply on RC1', sam keenor, 28 Mar 2025
  Dear Reviewer,
  We appreciate the time and effort that you have taken in evaluating our manuscript and for your constructive feedback, which we hope will help us improve the clarity and robustness of our work.
  Below, we provide responses to each of your comments and outline what potential actions we suggest.
  
  The methods were not introduced clearly, making it difficult to judge the results.
  
  What were the management practices (fertiliser, tillage, etc.) used for conventional farming?
  
  What were the grazing cover crops planted in the alleyways?
  
  What were the amount and nutrient content of the compost tea and organic fertiliser?
  
  Which year was the soil sample collected?
  
  This section initially opens with a brief outline of the field experiment, detailing the location and layout of the field sites used in the investigation, including a figure of field history. Broad information as to cultivation and management activities is provided for both the blackcurrant crop and the arable control field. The use of both the diverse grazing cover crop, and summer fallow planted into the fields during the investigation, as standard seed mixes were noted, however the specific species which made up these mixes were not explicitly stated.
  Given the soil physical property-centric focus of this investigation, we do not feel that the quantity and nutrient content of the compost tea organic fertilisers are important to the wider story told in this manuscript and would have minimal influence on soil aggregate stability and carbon storage. These additions were infrequent, and their carbon flux contribution would be insignificant when compared with the plant biomass carbon inputs.
  Year of soil sampling was omitted as an oversight – this can be clarified.
  As such we suggest the following actions:
  Provide further clarity as to cultivation depth and broader management in the arable control field
  
  If deemed necessary by the editor, include further information regarding diverse grazing and summer fallow mixes
  
  Include reference to the year and time of sampling within the methods section
  
  There were many different soil carbon types in this manuscript (e.g. labile and recalcitrant carbon, occluded carbon, stabilised carbon, WSA, and NWSA). The authors failed to separate carbon types clearly, It is hard to digest so many different terminologies.
  
  We recognise that the terminology may be complex. However, we believe we have clearly defined each soil carbon stock and its origin throughout the manuscript. To further improve clarity, we propose including a supplementary table that outlines and defines each type of carbon and its physical associations discussed in the study.
  As such we suggest the following action:
  An additional table within the supplementary section can be included which details each type of carbon discussed in the investigation and their origins. A glossary of terms if you will.
  
  Many parts in the Results and Discussion section were not about “results and discussion” but were about “background”, which should be moved to the introduction section
  
  In addition to adding a level of further interest and value to the manuscript, this content helps to define and delineate the types of carbon pools being discussed (raised as an issue above). While we believe that the background information within the Results and Discussion section is essential to contextualising the findings, we acknowledge the need to better balance content between sections.
  In revising the manuscript we will seek to better balance this content between the introduction and discussion sections.
  We suggest the following action:
  Balance the content between the introduction and discussion sections, being mindful of content that may be repetitive or inappropriately placed within the discussion section.
  
  More information about the field design and methods should be included in the abstract, to make it easier for readers to know the meaning of “alley soil and bush soil”.
  
  Due to word count constraints, we were limited in the level of detail provided in the abstract. However, we acknowledge the potential need for greater clarity in this regard and would be willing to add a concise acknowledgement of alley and bush soils.
  As such we suggest the following action:
  Following the Editor’s advice regarding word count and section length, we can seek to add a short statement detailing the alley/bush soils and their relevance within the research.
  
  The authors stated that “This research provides valuable insights into the mechanisms of soil carbon stabilisation under regenerative agriculture practices.” However, I am not sure if they really unravel the mechanisms because they did not measure other soil variables (e.g. soil nutrient content and soil microbial parameters), which can be used to explain the results.
  
  While we appreciate this concern, soil nutrient content and soil biodiversity parameters were not the focus of this investigation. Our research reveals the collective outcome of a switch to regenerative agricultural principles as a soft fruit crop establishes. We feel that the insights regarding labile vs recalcitrant carbon stocks and their association with stable and non-stable aggregates does amount to appreciation of the mechanisms of soil carbon protection.
  We acknowledge that there are, of course, a myriad of causal factors at play that may influence the outcome. As such, we highlight that this is a field experiment and control of parameters is challenging. Even if for example nutrient content and soil microbial parameters were elucidated, we do not feel such information would support causality of the outcome.
  Additionally, measuring these additional variables such as soil microbiome dynamics would require near-continuous monitoring, which was beyond the scope of our research.
  We suggest no further action.
  
  The “Results and Discussion” section was mainly introducing background and describing the results. Instead, more deep discussions and measurements should be included to explain the interesting results (e.g. when compared to control treatment, why did total carbon stock decrease in the short term and then increase to the similar level like the control treatment in the long term). Does this mean regenerative agriculture must be conducted for a long time? Otherwise, croplands would lose soil carbon?
  
  We feel that the results and discussion section answers adequately these points within the constraints of the experimental investigation that was undertaken. We emphasise that the observed trends suggest that long-term management is required to achieve significant recalcitrant carbon gains.
  We feel that the text in its present form already addresses these points, however, we welcome specific guidance from the reviewer on any aspects requiring further clarification.
  We suggest no further action, pending a point of clarification from the reviewer.
  
  Once again, we sincerely thank you for your insights and your time.
  Kind regards,
  Sam G. Keenor, and Brian J. Reid
  
  Citation: https://doi.org/10.5194/egusphere-2024-4029-AC1
RC2:
'Comment on egusphere-2024-4029', Anonymous Referee #2, 03 Mar 2025

This paper explores the impact of regenerative management on soil aggregates and carbon pools. The samples originate from blackcurrant fields (bushes and alleys) with duration from 1 to 7 years, and a control plot still under conventional management. Many analyses were conducted (soil bulk density, stable and non-stable aggregate contents, soil carbon contents and stocks, and carbon stability). The main observation is a shift towards more stable aggregates and an increase of the labile carbon pool under regenerative management, with little effect on the recalcitrant pool.
The aim of this research and the results are interesting. Although I think the article is not acceptable at this stage, I would recommend resubmitting the manuscript after signficant changes and thorough proofreading to enhance the work that has been conducted. Some of the elements that should be reconsidered follow.
The article lacks clarity regarding crucial details such as the experimental set-up (number of fields, number of samples and replicates, etc.) and the presentation of results. Figure 1 from the supplement should be included in the article to provide a clearer picture of the set-up; a map of the experiment could also bring some clarity. It would also be interesting to discuss the prior presence of blackcurrant fields in years 5 and even 3: did the introduction of new plants lead to a return to a state of conventional management, thereby 'resetting' the regenerative agriculture counter to 0 years?
The results should be discussed further and the text should be better organized between what is general introduction of the topic, results, and discussion. Even if results and discussion can be addressed altogether, they should be clarified and strengthened.
Overall, the article is not easily readable. Check for typos, missing verbs or parts (abrupt end L.631 for example). Indications regarding p-values appear repeatedly throughout the text (as you defined two levels of confidence) without really bringing relevant information; the p-values could appear on your graphs but not in text for instance, with only mentions to 'significant' or 'not significant'. Also, a genuine thought: could some results be non significant due to the small numbers of samples? Would other tests be better at handling this issue?

Citation: https://doi.org/10.5194/egusphere-2024-4029-RC2
- AC2:
  'Reply on RC2', sam keenor, 28 Mar 2025
  Dear Reviewer,
  We appreciate the time and effort that you have taken in evaluating our manuscript and for your constructive feedback, which we hope will help us improve the clarity and robustness of our work.
  Below, we provide responses to each of your comments and outline what potential actions we suggest.
  
  The article lacks clarity regarding crucial details such as the experimental set-up (number of fields, number of samples and replicates, etc.) and the presentation of results.
  
  This section initially opens with a brief outline of the field experiment, detailing the location and layout of the field sites used in the investigation, including a figure of field history. As such, the number of fields sampled during this investigation in addition to the number of samples are explicitly stated within the methodology sections.
  However, in line with comments from both reviewers, we accept that there could be increased clarity regarding some aspects of the methodology section of the manuscript.
  As such we suggest the following actions:
  Provide additional clarity and changes to the methodology, with focus placed specifically on clarifying the number of replicates and total samples taken.
  
  Provide further clarity as to cultivation depth and broader management in the arable control field
  
  If deemed necessary by the editor, include further information regarding diverse grazing and summer fallow mixes
  
  Include reference to the year and time of sampling within the methods section
  
  Figure 1 from the supplement should be included in the article to provide a clearer picture of the set-up; a map of the experiment could also bring some clarity.
  
  We agree that Figure 1 could be made as part of the main article, a map of the experiment can also be produced if required by the editor.
  We suggest the following actions:
  Move Figure 1 to an appropriate place within the main article
  
  Pending the Editors judgement, produce a map as required
  
  It would also be interesting to discuss the prior presence of blackcurrant fields in years 5 and even 3: did the introduction of new plants lead to a return to a state of conventional management, thereby 'resetting' the regenerative agriculture counter to 0 years?
  
  We agree that there is a point of interest here. Given the methods of establishment – between the cultivation of a new crop of blackcurrant bushes (i.e. field Y5) the field undergoes a total cultivation essentially ‘resetting’ the field as the reviewer identifies. The methodology can be updated to clarify that total cultivation between plantings effectively resets the soil conditions.
  We suggest the following actions:
  Clarify the ‘reset’ of the soil following total cultivation between plantings of blackcurrants.
  
  Results should be discussed further, and the text should be better organised between what is general introduction of the topic, results, and discussion. Even if results and discussion can be addressed altogether, they should be clarified and strengthened
  
  While we believe that the background information within the Results and Discussion section is essential to contextualising the findings, we acknowledge the need to better balance content between sections.
  In revising the manuscript, we will seek to better balance this content between the introduction and discussion sections.
  We suggest the following action:
  Balance the content between the introduction and discussion sections, being mindful of content that may be repetitive or inappropriately placed within the discussion section.
  
  Overall, the article is not easily readable. Check for typos, missing verbs or parts, abrupt end L.631 for example
  
  We will provide additional grammar and spell checks when reviewing potential changes to the article. We agree there is an error in L631.
  As such we suggest the following action:
  Review grammar, spelling and readability of the text.
  
  Indications regarding p-values appear repeatedly throughout the text (as you defined two levels of confidence) without really bringing relevant information; the p-values could appear on your graphs but not in text for instance, with only mentions to 'significant' or 'not significant'.
  
  We would be happy to make this amendment and move the p-values to figures instead of being repeatedly stated in the text. We highlight that such an approach in the case of the stacked histograms may be too confusing, to be included on the figure itself alongside the other statistical notations, however accommodation of P-values could be made within figure legends for clarity. We will act as advised by the editor.
  We suggest the following action
  Follow advice of the editor and make the necessary alterations to statistics reporting in the text and on the Figures.
  
  Could some results be non-significant due to the small numbers of samples? Would other tests be better at handling this issue?
  
  For statistical analysis, a one way ANOVA with post-hoc testing was used to determine significance at a 95% confidence level, alongside 2-tailed T-tests – both are robust and standard statistical tests for analysis of this type of data. We do not believe that altering the statistical testing methodology would provide any benefit to the analysis. Furthermore, the sample numbers are, as they are. There is no prospect to generate more measurements.
  We suggest no further action.
  
  Once again, we sincerely thank you for your insights and your time.
  Kind regards,
  Sam G Keenor, and Brian J. Reid
  
  Citation: https://doi.org/10.5194/egusphere-2024-4029-AC2

Peer review completion

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

ED: Reconsider after major revisions (further review by editor and referees) (20 Apr 2025) by Katerina Georgiou

AR by sam keenor on behalf of the Authors (16 Jun 2025) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (04 Jul 2025) by Katerina Georgiou

RR by Guusje Koorneef (18 Jul 2025)

Suggestions for revision or reasons for rejection

The authors pose an interesting and relevant question in the light of climate change and growing interest for and implementation of regenerative farming. The system approach used for the design of the field experiment is interesting, as well as choosing blackcurrants as crop that I have not often encountered.
However, I see serious flaws in the methodological design that to my opinion hamper answering the posed research question, i.e., whether transition from conventional to regenerative farming leads to increased SOC sequestration. This has to do with 5 issues:
1) Authors focus on assessing SOC stocks in the field, which is the amount of SOC stored across the soil profile at minimally 30 cm depth per unit area, according to the IPCC guidelines (Penman et al 2003, “Good practice guidance for land use, land-use change and forestry”). Especially in the case of reduced tillage practices as part of the transition towards regenerative farming, assessing SOC stocks at multiple depths is relevant. Reduced tillage is known to ‘concentrate’ SOC in the upper soil layers, as this SOC is no longer mixed with lower soil layers. Therefore, SOC contents at lower soil layers usually is ‘diluted’ with reduced tillage, so that SOC stocks across the entire soil profile may not be that different afterall. Therefore, SOC stocks should have been measured at lowers depths, preferable even lower than 30 cm as subsoil processes are very different from topsoil processes (Rovira et al, 2022, CATENA). In this study, SOC is only measured at 0-7.5cm depth (line 144, and I think this is insufficient given the information above.
2) Authors mention multiple times the carbon sequestration potential when interpreting the results (e.g. lines 20, 26, 563). However, it is important to distinguish carbon sequestration (truly capturing CO2 from air into soil) vs. carbon rereallocation (moving carbon from one field to another), see Leifeld et al 2013, PNAS. To assess whether the study design can inform the carbon sequestration potential, it is important to know the organic matter inputs to the field. However, in section 2.1 I do not see systematic information on the type and amount of fertilization and organic matter inputs. The use of organic fertilizer and compost tea for regenerative farming is mentioned for regenerative farming, and stubble re-incorporation for conventional farming. However, it makes the reader wonder on the fertilization applied in the conventional field (no info given), and crop residue management in the regenerative field (no info given). Given the application of organic fertilizer in the regenerative fields, I would say that at least part of the increased SOC contents derive from carbon reallocation rather than sequestration.
Moreover, I also miss more information given the tillage operations (e.g. depth), given its importance for SOC transfer across the soil profile.
3) The concept of biochemical recalcitrance of carbon is outdated, except for pyrogenic carbon (Schmidt et al, 2011, Nature), unless argued otherwise. However, this concept is introduced in lines 67-68 without reference, and interpreting the TGA results as currently done (labile vs. recalcitrant SOC) is only supported by one reference without theoretical explanation. It also makes me wonder whether physically protected SOC within aggregates is part of this recalcitrant fraction? I wonder why it is not chosen to use the t50 of TGA that accounts for the continuum of soil carbon stability (Lehmann et al 2015, Nature)?
I believe deviating from these common concepts and instead use a simplified 2-pool classification of labile vs recalcitrant SOC needs to be much more substantiated, and based on theoretical stabilization mechanisms of SOC, as is done for the distinction of POM-MAOM by size or density fractionation by Lavallee and Cotrufo.
4) Section 2.6: Given the changed tillage regime that potentially alters bulk density, I believe that not the fixed depth method but the equivalent soil mass method should have been used to calculate SOC stocks from SOC content (Von Haden et al 2018, GCB)
5) Section 2.7: what about meeting the assumptions for these models, have they been checked, and can hence the statistical results be trusted?
Based on the observations, I recommend to reject this manuscript in its current form for publication. However, I do believe that the experiment and results are super interesting and can be used if the story is not focused on assessing SOC stocks and carbon sequestration, but on assessing the changes in SOC quality and aggregate stability in top soil after a system transformation towards regenerative farming, acknowledging the continuum of SOC stability and providing more quantitative management information relevant for SOC characteristics.
For now, I mostly read the abstract, introduction and materials and methods. If the manuscript is continued with publication in its current form, I can review results and discussion in more detail if requested (after my holidays… 😊).

Hide

RR by Anonymous Referee #2 (17 Aug 2025)

Suggestions for revision or reasons for rejection

This paper explores the impact of regenerative management on soil aggregates and carbon pools in blackcurrant fields with different durations ranging from 1 to 7 years. The control is a nearby plot under conventional management. The main observation is an increase of the water stable aggregates and the labile carbon pool under regenerative management. Little effect is observed on the recalcitrant pool, although this may be a result of the ‘short’ duration of the experiment.

The subject and its implications are important and well detailed. The study site and the methods are described. Results and Discussion are presented together. A lot of work has been conducted here, thus many results are presented, which may be detrimental to the clarity of the paper: while explained, the numerous notions (WSA/NWSA, labile/recalcitrant, aggregates, thermal stability in aggregates, bush/alley soils, several years…), make the article uneasy to read.
I suggest some data appear only in Supplementary Information (e.g. bush soils detailed in the paper and alley soils only in SI). You could also consider making a shorter article, focusing more on certain results, and backing it up by a data paper for the rest of the results.

Here are a few specific remarks:
L24: it would be good to reformulate this sentence, as the ‘proportion of NWSA : WSA’ is not really ‘increasing’ (only the % of WSA is indeed increasing).
L71: non verbal sentence, I guess it is rather the end of the previous sentence. Same for L81.
L88: soil properties regulation
L198: ‘that which passed through the 63 µm sieve after the HMP treatment’
L222: what is expected between 700 and 1.000°C? (Or – why not stop heating before 1.000°C?)
L252: any idea why the lowest global SBD is in year 3 treatment and not year 7?
L262: while it does not fully change the overall observation, I would not say ‘generally decrease with time’
L318: there is no year 2 soil, I guess it is 3.
L342-347: So, recalcitrant carbon content increases with time, but not significantly? This paragraph is not totally clear.
L352: if you mention the increase for year 7, you should also mention the one for year 5 as you talk about both years. Same for L359.
General: check for punctuation, spelling and grammar. Several non-verbal sentences are still found along the text.

Referee Report: PDF

Hide

ED: Publish subject to minor revisions (review by editor) (14 Sep 2025) by Katerina Georgiou

Dear authors,

Thank you for carefully responding to the reviewers comments. In general, the manuscript has improved and represents a large amount of work. The results are broadly interesting, but the manuscript remains convoluted and difficult to follow at times, as Reviewer #2 points out. Consider this carefully and move surplus discussion to the supplement to aid readability. Also, proof-read again for sentence fragments and errant sentences outside of paragraphs (some examples given below). Please go ahead with addressing the reviewer comments and replying to any points for discussion.

A few additional comments are included below:

- Remove the word “associated” from NWSA and WSA in Table S1, to match how it is used throughout the text. Also, consider adding the word “thermally” to labile and recalcitrant carbon to specify the operational definition. In fact, potential caveats of this operational definition and other mechanisms governing “recalcitrance” (as brought up by Reviewer #3 and reviewed in Schmidt et al. Nature 2011) should be mentioned briefly when the method is first introduced (line 134). This is all nicely done in the title and text of section 3.3, but could be added briefly to the introduction and throughout.

- Spell out abbreviations in figure captions to make it easier for figures to be understood on their own.

- Line 46 – consider replacing the semi-colon with a comma or rewording into two sentences.

- Line 59 - combine this sentence with the paragraph below and remove first comma. Also, consider splitting the following paragraph which is quite long.

- Line 132 – the word respectively is not needed here, since it seems that it refers to the bush and alley soils which are specified in their respective places in the sentence. Check and reword for clarity.

Line 267 - combine sentence with previous paragraph. In fact, there are many spots like this throughout the text where a sentence or two stand as their own paragraph (e.g., line 362, 393, 440, 443, 447, 452, etc). This disrupts the flow and connections between findings. Please find and rearrange as needed.

Thank you,
Katerina

Hide

AR by sam keenor on behalf of the Authors (03 Oct 2025) Author's response Author's tracked changes Manuscript

ED: Publish as is (19 Oct 2025) by Katerina Georgiou

ED: Publish as is (19 Oct 2025) by Rémi Cardinael (Executive editor)

AR by sam keenor on behalf of the Authors (22 Oct 2025) Author's response Manuscript

Short summary

Regenerative soil management reduced bulk density, increased carbon stocks, and influenced aggregate stability with time. Soils were observed to become proportionally more enriched in stable aggregates and store more carbon with time. Stable aggregates provided physical protection (occlusion) to labile carbon stocks. When considered alongside recalcitrant carbon, these carbon stabilisation mechanisms may provide further opportunities to deliver robust long term carbon sequestration outcomes.