Review Reports - Soil Carbon Content in Areas with Different Land Uses and Vegetation Cover in the Cerrado–Amazon Transition, Mato Grosso, Brazil

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript presents an analysis of soil carbon content in two watersheds of the Cerrado and Brazilian Amazon regions. The approach includes analysis at different soil depths, different land covers (land uses), and different watershed zones, with robust and representative sampling of the area. The study of carbon sequestration and its dynamic changes across land uses in different biomes is highly relevant for understanding the role of soils in the context of global environmental change, making the manuscript current and pertinent. The manuscript is well structured, and the methods are adequate. My main observation is directed toward the sampling design and the presentation of results. When conducting studies at a regional scale, it is important to address soil natural spatial variability in the statistical design. Although the manuscript presents the soil classes identified in the region, their delimitation is not clearly considered in either the sampling design or the analysis of results. In addition, the association of main effects and interactions between the variables of the 3x3 design (coverage, depth, zone) is not consistent in the presentation of results. Overall, the study provides findings of interest to the scientific community and aligns with the scope and readership of Soil Systems. Before recommending publication, several revisions are necessary:

Abstract:

Lines 26, 28. Use standard notation for density. “Pd” is confusing.
Lines 29–31. Add uncertainty.
Specify/unify nomenclature: C content / total C / Soil C / C stock / CC. Use a consistent term throughout the manuscript.

Introduction: The introduction needs improvement to highlight the relevance and novelty of the study. Strengthen the introduction by expanding the global perspective on land-use change and carbon sequestration in different ecosystems, then focus on the Cerrado and Amazon biomes.

Lines 39–48. It is recommended to begin by presenting the regulatory function of soils in the carbon cycle to emphasize the global importance of studying this phenomenon in various ecosystems under global environmental change.
It is recommended to present the relationship between land-cover change and carbon sequestration from the perspective of planetary boundaries, one of the most up-to-date sustainability frameworks linking soil conservation with its role in climate change mitigation.
Reference: Steffen, W., Richardson, K., Rockström, J., Cornell, S. E., Fetzer, I., Bennett, E. M., … & Sörlin, S. (2015). Planetary boundaries: Guiding human development on a changing planet. Science, 347(6223), 1259855.
State a research hypothesis. Although the importance of associating physical soil indicators with carbon storage is mentioned, there is no clear formulation of the scientific problem in the regional context. Why are differences expected between upper, middle, and lower watershed zones, rather than considering variations associated with soil type or landscape unit? Expand the problem statement.

Methods: In general, the methodology is adequate and well described. Review the following comments:

Lines 84–88. Indicate to which system the analyzed sub-watersheds belong and their final outflow.
Line 90. Define “Aw.”
Line 101. Not all international readers are familiar with the Brazilian classification system. Consider providing equivalence between Brazilian soil taxonomy and USDA or WRB systems.
Provide information on the typical sequence of soil horizons. Was the depth of the surface horizon measured in sampling sites?
Lines 111–116. Provide historical information on land uses: dynamics of agricultural frontier expansion and the age of crops and pastures established in the area.
Line 121. Review translation of (Safrinha).
Line 159. Define TFSA.
Line 164. Provide equipment characteristics and details of calibration and quality-control procedures for elemental analysis (linearity, blanks, memory effect control).
Lines 168–174. Formula 1: Explain the “10” in the denominator. Review acronym definitions and use symbols consistently with earlier sections. Ds (Bulk).

Results: The presentation and interpretation of results should be improved. Furthermore, there is a lack of consistency in the presentation of results with regard to the relationship between variables established in the research objective

Why did the statistical design not account for soil taxonomic units? Physical properties are clearly influenced by parent material and topography. For the upper, middle, and lower watershed zones, average soil texture values are presented based on sampling points from different taxonomic units, which causes confusion when assessing the effect of land cover and watershed zone. Did the authors explore comparing properties across land covers within individual soil taxonomic units?
Ensure consistency between the study objective and the variables analyzed for carbon sequestration. Although a 3×3 design is explained, results presentation seems to conflate depth effects, land cover, and changes between the two micro-watersheds (upper-middle-lower zones). This must be clarified starting from the research hypothesis to properly structure the results and statistical analyses around the key variables expected to differ.
Present results with standard deviation (Mean ± SD), both in text and in Tables 1, 2, 3. Without this, variation levels remain unclear throughout the manuscript. A scatterplot would also help.
Review table footnotes (Tables 1 and 2), and explain the meaning of capital letters.
What is the explanation for the result of 11.57 g kg C in the surface of cropland in the upper part of the Renato River (Table 3)?
Lines 290–298. Specify the number of samples used in the correlations. It is unclear whether correlations were conducted separately by watershed or on pooled data. Different trends are mentioned for Caiabi and Renato Rivers, but Table 4 coefficients combine both results. Did the authors explore correlations grouped by land covers?
It is unclear why PCA is presented grouped by depths and by watershed. Based on the research objective (Lines 77–81), analysis by land cover would be more meaningful for providing practical recommendations on soil management and conservation. PCA should be reassessed. Explore PCA approaches involving categorical variables to unify the analysis and show the influence of depth, micro-watershed, zone, and soil taxonomic class.

Discussion: The discussion needs improvement.

Lines 460–475. Expand discussion with information on land-use change trends. What is the current dynamic of these activities in the region? Are any expanding or declining? There is extensive literature supporting how shifts between pasture and forest, forest and cropland, cropland and pasture, etc., affect carbon reserves. Practical implications and management recommendations could therefore be more specific, not only regarding management practices but also in terms of land-use regulation and planning.
After revising results, strengthen the discussion on carbon stocks. Discuss separately and hierarchically the effects of region, soil taxonomic class, land cover, depth, and interactions with other soil properties.
Also, consider the study’s limitations. Chemical and biological variables also influence carbon reserves and physical processes. Although the study’s objective focuses on physical properties, it is important to discuss possible effects associated with biological activity and chemical processes that significantly affect carbon content changes across depths and land covers.

Conclusion:
The conclusion should be reformulated, as it does not provide a concrete answer to the study’s objective. Which variable in the 3×3 design has the greatest influence on carbon content, and what are the interactions among them?

Author Response

Review Report Form - Reviewer 1

Comments and Suggestions for Authors

RESPONSE: We appreciate the contributions and suggestions of all reviewers. Their comments and evaluations reinforced the credibility and scientific quality of our article. We want to inform you that the manuscript has undergone some changes to address all suggestions. We are very grateful for the reviewers' comments on the spatial extrapolations and representativeness of the study, which demonstrate the originality and applicability of the results for soil conservation in areas of agricultural expansion.

Abstract:

Lines 26, 28. Use standard notation for density. “Pd” is confusing.

RESPONSE: Corrected

Specify/unify nomenclature: C content / total C / Soil C / C stock / CC. Use a consistent term throughout the manuscript. -

RESPONSE: The nomenclature in the abstract and throughout the article was corrected, standardizing the acronyms and terms.

Introduction:

The introduction needs improvement to highlight the relevance and novelty of the study. Strengthen the introduction by expanding the global perspective on land-use change and carbon sequestration in different ecosystems, then focus on the Cerrado and Amazon biomes.

Lines 39–48. It is recommended to begin by presenting the regulatory function of soils in the carbon cycle to emphasize the global importance of studying this phenomenon in various ecosystems under global environmental change.

RESPONSE: The first paragraph of the Introduction (lines 39-48) was inserted to address this comment. We appreciate the suggestion, as it allowed us to approach the writing in a deductive format. In addition, new current references were added that show the interrelation with climate change.

State a research hypothesis. Although the importance of associating physical soil indicators with carbon storage is mentioned, there is no clear formulation of the scientific problem in the regional context. Why are differences expected between upper, middle, and lower watershed zones, rather than considering variations associated with soil type or landscape unit? Expand the problem statement.

RESPONSE: The scientific hypothesis was included in the last paragraph of the introduction (lines 90-92) to support the objectives established in the article and the logical sequence of the other textual items.

Methods:

In general, the methodology is adequate and well described. Review the following comments:

RESPONSE: Lines 84–88. Indicate to which system the analyzed sub-watersheds belong and their final outflow.

RESPONSE: Line 101-102: … The basin belongs to the larger Amazon basin and features predominant vegetation of the Cerrado …

RESPONSE: Lines 125-130: The Caiabi River sub-basin has a drainage area of approximately 493 km² and a mean flow of 9,4 m³ s^-1, covered with 31% natural Cerrado-Amazon transition vegetation, 60% monoculture (soybean-corn succession) and 8% pasture. The Renato River sub-basin has a drainage area of approximately 1,336 km² and a mean flow of 16,3 m³ s-1, which is occupied by 69% Amazon Forest subject to forest management and with significant areas converted for pasture (19%) and crops (11.4%) [28] (Figure 2).

Line 90. Define “Aw.”

RESPONSE: Line 104-108: The predominant climate in the study region is Aw, A – very humid; w – summer rains, considered a Tropical Savanna Climate (hot and humid tropical) with average monthly temperatures between 24 and 27 °C and two well-defined water seasons: dry oc-curs in autumn/winter (May to September) and rainy occurs in spring/summer (October to April), with annual rainfall between 1,800 and 2,200 mm [27].

Line 101. Not all international readers are familiar with the Brazilian classification system. Consider providing equivalence between Brazilian soil taxonomy and USDA or WRB systems.

RESPONSE: Corrected in accordance with “Keys to Soil Taxonomy. (2014). United States Department of Agriculture, Natural Resources Conservation Service. 372p.” Standardization was carried out throughout the entire text.

Provide information on the typical sequence of soil horizons. Was the depth of the surface horizon measured in sampling sites?

RESPONSE: The sequences and evaluations of soil horizons in the two hydrographic sub-basins were carried out by opening 9 soil classification profiles (up to 2.0 m deep) in each sub-basin, totaling 18 profiles; in this case, 3 profiles per region of each sub-basin were considered. The descriptions and classification characteristics of the 18 profiles were presented by Alves et al. (2022) - https://doi.org/10.31413/nativa.v10i3.14192 (reference [32]); these soil profile classification sampling points (horizons) were at most 10 m away from the sampling points (trenches) for collecting disturbed and undisturbed soil samples, for defining physical and carbon attributes.

Lines 111–116. Provide historical information on land uses: dynamics of agricultural frontier expansion and the age of crops and pastures established in the area.

RESPONSE: Information and details entered on lines 135-147

Line 121. Review translation of (Safrinha).

RESPONSE: Line 148: (harvest and second harvest)

Line 159. Define TFSA.

RESPONSE: Line 196: (air-dried fine soil)

Line 164. Provide equipment characteristics and details of calibration and quality-control procedures for elemental analysis (linearity, blanks, memory effect control).

RESPONSE: Information and details were inserted between lines 201 and 208.

Lines 168–174. Formula 1: Explain the “10” in the denominator. Review acronym definitions and use symbols consistently with earlier sections. Ds (Bulk).

RESPONSE: Information inserted between lines 218 and 219

Results:

The presentation and interpretation of results should be improved. Furthermore, there is a lack of consistency in the presentation of results with regard to the relationship between variables established in the research objective

Why did the statistical design not account for soil taxonomic units? Physical properties are clearly influenced by parent material and topography. For the upper, middle, and lower watershed zones, average soil texture values are presented based on sampling points from different taxonomic units, which causes confusion when assessing the effect of land cover and watershed zone. Did the authors explore comparing properties across land covers within individual soil taxonomic units?

RESPONSE: We appreciate your comment and concern. Although there are differences in the physical characteristics of the soils, and these are indeed directly related to pedogenetic processes and influenced by formation factors, specifically in the case of these two hydrographic sub-basins, the same taxonomic group of soils is observed: Oxisols. Furthermore, the small textural differences along the hypsometry of the sub-basins (between the regions) result from transport and centuries-old geological sedimentation, mainly in the estuary regions near the Teles Pires River. Therefore, it is believed that the greatest influence on carbon stock is related to the type of vegetation cover (LULC), combined with its position in relation to each hydrographic sub-basin.

Ensure consistency between the study objective and the variables analyzed for carbon sequestration. Although a 3×3 design is explained, results presentation seems to conflate depth effects, land cover, and changes between the two micro-watersheds (upper-middle-lower zones). This must be clarified starting from the research hypothesis to properly structure the results and statistical analyses around the key variables expected to differ.

RESPONSE: Inserted between lines 87 and 92, highlighting the relationship between soil carbon stocks in basin regions and land uses.

Present results with standard deviation (Mean ± SD), both in text and in Tables 1, 2, 3. Without this, variation levels remain unclear throughout the manuscript. A scatterplot would also help.

RESPONSE: This article is a continuation of the same research project, whose initial article was published in Soil Systems, using the same analysis pattern, for the context of the two hydrographic sub-basins (it can be viewed at https://doi.org/10.3390/soilsystems8010031).

Due to the large number of variables and analyses, including SD along with the mean values in Tables 1 and 2, this would create an additional difficulty in understanding the possible differences, and the tables would cease to be self-explanatory. We agree with the importance of presenting SD values to understand the variability of the study; however, this could be done as a supplementary document or appendix.

Review table footnotes (Tables 1 and 2), and explain the meaning of capital letters.

RESPONSE: Clarifications are provided in the footnotes to the two tables.

Different uppercase letters in a column (when analyzed in different sub-basin regions for the same land use and soil depth) differ significantly from each other in the non-parametric Kruskal-Wallis test (p < 0.05). Different lowercase letters in a row (when analyzed in different land uses in the same region and soil depth) differ significantly from each other in the non-parametric Kruskal-Wallis test (p < 0.05). Mean values that do not contain letters do not differ from each other.

What is the explanation for the result of 11.57 g kg C in the surface of cropland in the upper part of the Renato River (Table 3)?

RESPONSE: Because it is also a sandy region, surface carbon may have migrated through the profile to the subsurface layers, especially to the 0.10-0.20 m layer, as occurs with clays in eluviation and illuviation processes, which is also reflected in the reduction in the 0.20 to 0.40 m layer.

Lines 290–298. Specify the number of samples used in the correlations. It is unclear whether correlations were conducted separately by watershed or on pooled data. Different trends are mentioned for Caiabi and Renato Rivers, but Table 4 coefficients combine both results. Did the authors explore correlations grouped by land covers?

The following sentences were excluded from the summary:

“The data were subjected to Spearman's correlation analysis and the Kruskal-Wallis test at 5% probability”.

“The adoption of conservation practices and government incentives can improve soil C.”

RESPONSE: Details were inserted between lines 162 and 174, highlighting the number of samples evaluated in each region and land cover in each sub-basin. For Spearman's correlations, no separation of regions and land covers was performed; that is, all data were grouped by sub-watershed, aiming to verify general correlations, since it is the same soil class. However, significant differences were observed for some physical soil attributes (Tables 1 and 2). Several significant correlations are noted between attributes, regardless of layer depth, land cover, and sub-watershed region. Increasing the degree of stratification of the correlations does not necessarily improve the correlations, as greater dispersion between independent variables may be obtained for the same range of dependent variable in the correlation. We understand that the general grouping performed by sub-watersheds already allows establishing regional relationships and more practical applications at the field level.

It is unclear why PCA is presented grouped by depths and by watershed. Based on the research objective (Lines 77–81), analysis by land cover would be more meaningful for providing practical recommendations on soil management and conservation. PCA should be reassessed. Explore PCA approaches involving categorical variables to unify the analysis and show the influence of depth, micro-watershed, zone, and soil taxonomic class.

We understand that the analyses performed are coherent and allow us to achieve the main objective of PCA, which is to simplify/group the influences of variables into "principal components," in this case, considering representativeness greater than 75% (combined) of PC1 and PC2 for the layers and sub-basins; altering this form of presentation would open up a larger number of figures (by region or by use) and this would not add representativeness much greater than 75% (the minimum required for biplot-type applications).

Discussion:

The discussion needs improvement.

Lines 460–475. Expand discussion with information on land-use change trends. What is the current dynamic of these activities in the region? Are any expanding or declining? There is extensive literature supporting how shifts between pasture and forest, forest and cropland, cropland and pasture, etc., affect carbon reserves. Practical implications and management recommendations could therefore be more specific, not only regarding management practices but also in terms of land-use regulation and planning.

Lines 111–116. Provide historical information on land uses: dynamics of agricultural frontier expansion and the age of crops and pastures established in the area.

RESPONSE: Text inserted between lines 135 and 147 to accommodate this comment.

After revising results, strengthen the discussion on carbon stocks. Discuss separately and hierarchically the effects of region, soil taxonomic class, land cover, depth, and interactions with other soil properties.

RESPONSE: We appreciate the comment and suggestion. A thorough review of the results was conducted. We understand that the structure of the discussion and the form of presentation follow a logical sequence that allows for regional contextualization and its comparison with larger spatial scales.

Also, consider the study’s limitations. Chemical and biological variables also influence carbon reserves and physical processes. Although the study’s objective focuses on physical properties, it is important to discuss possible effects associated with biological activity and chemical processes that significantly affect carbon content changes across depths and land covers.

RESPONSE: Two paragraphs were added to the Discussion to address this comment (lines 572 to 589).

Conclusion:

The conclusion should be reformulated, as it does not provide a concrete answer to the study’s objective. Which variable in the 3×3 design has the greatest influence on carbon content, and what are the interactions among them?

RESPONSE: The conclusions were rewritten to answer the objectives in a simple, clear, and direct way.

Reviewer 2 Report

Comments and Suggestions for Authors

This is an interesrtingmanuscript for the readers of the journal and i in the score of the journal, aiming to invetigateffect of lan use on soil propertiessnd stock st different psrt of two basins

Consider to evaluate effect of region x elevation x land use on soil C

Introduction

Needs streaming to formulate the objective properly. Please also try to concentrate on similar regions (climate soil types elevation soil textue, cover type etc

Line 70: and soil structure and quality indexes could be becused on the bais of vailable data.

Line 78: from point of what?

Material and methods

Clear sample number should be provided: by the basin, region, land use, soil type.

Multicolor analysis should be provided (basin, region, land use and their interaction effect on soil properties and C stock. Stepwise analysis can help to find order of soil properties that effect on C content.

Results

Line221-225 zhould be in the discussion section

For each each basin Spearman correlation should be done.

Discussion

C accumulation and loss mechanisms for each in each region or land use and soil type should be clearer

Duration of each land use should be provided

Conclusion

Cosider to add practical outcomes

Author Response

Comments and Suggestions for Authors

This is an interesting manuscript for the readers of the journal and in the scope of the journal, aiming to investigate the effect of lan use on soil properties in different parts of two basins.

Consider evaluating the effect of region x elevation x land use on soil C

RESPONSE: We appreciate your feedback on the article and your valuable contributions. This is an excellent consideration; however, in this specific article, it is believed that there is no need to test the effect of altitude in relation to other factors on soil carbon stock, since, according to Figure 1, the altitude variations are relatively small and were considered in the delimitation of the watershed regions (source, middle, and mouth). We also added a sentence (lines 171-172) to clarify the sampling at midpoints of the toposequence.

Introduction

Needs streaming to formulate the objective properly. Please also try to concentrate on similar regions (climate soil types elevation soil textue, cover type etc

Line 70: and soil structure and quality indexes could be becused on the bais of vailable data.

RESPONSE: Soil structure and quality indices can be indirectly inferred based on the results/data obtained in this work; however, the soils in this region are highly weathered, with well-defined structures and, for the most part, granular in the diagnostic horizons; some articles already developed in agricultural areas in the Northern region of Mato Grosso (corresponding to Southern Amazonia) show this relationship:

https://doi.org/10.1016/j.geoderma.2020.114796

https://doi.org/10.1016/j.eja.2020.126090

Since we did not perform these structural analyses, specifically at the sampled points, we opted to make indications/comments about these possibilities, to avoid "speculation".

Line 78: from point of what?

RESPONSE: Corrected

Material and methods

Clear sample number should be provided: by the basin, region, land use, soil type.

RESPONSE: Clear identification of the number of trenches opened in each basin, region, and land use was inserted between lines 160 and 172. According to the work of Alves et al. (2023) - https://doi.org/10.31413/nativa.v10i3.14192, the soil classes in the different regions of each watershed are the same, with differences only in the averages for some physical-hydric attributes of the soils.

The APC results group the information independently of the effects of the interactions, being separated only according to the soil layer, since cultural and management practices are effective at depth and may show a greater relationship with possible changes in SOC and CS. Multicolor identification in this case would lose its effect.

Stepwise analysis is an automated regression technique that selects the most relevant predictor variables for a model by iteratively adding or removing variables based on statistical criteria (such as F-value or AIC) to find the best fit to the data, combining "forward" (addition) and "backward" (removal) methods. While useful for exploring large datasets and identifying important predictors, it can generate models that may not be generalizable or theoretical, and is often discouraged for research with clear hypotheses, such as that of this work. We did not use it in this article, since the generated database does not allow for external validation; therefore, it would be a very small quantity of sample data for calibration/validation. With considerable caution, and when we have access to Big Data on the chemical, physical, and biological attributes of the soils in this region, it will certainly be an excellent option for modeling and understanding the dynamics and effects on regional soils; however, always with validation.

Results

Line 221-225 should be in the discussion section

RESPONSE: We agree with the comment and have made the change (Lines 502-507)

For each basin Spearman correlation should be done.

RESPONSE: Although they are in the same table (Table 4), the correlation was performed for the two basins separately: Spearman correlation significant at 5% and 1% probability, respectively; orange color: Caiabi; green color: Renato.

Discussion

C accumulation and loss mechanisms for each in each region or land use and soil type should be clearer

RESPONSE: Some adjustments were made to the text, and a paragraph was also inserted to make the discussion (Lines 401-408)

Duration of each land use should be provided

RESPONSE: Text inserted to address this comment, between lines 133 and 145.

Conclusion

Consider adding practical outcomes

RESPONSE: We agree with the comment and have already made all the necessary changes to the article's conclusions.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I thank the authors for their clarifications and responses to the reviews. The new version of the manuscript satisfactorily incorporates the recommendations.

Author Response

We appreciate your collaboration in evaluating the article and especially your favorable opinion regarding its publication.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors mostly addressed all comments . their responce is ientifically sound and generally well justified, but two partial points need attention

1) No formal region × elevation × land use interaction test; 2) No explicit interaction modeling (despite justification). These are not fatal and might be accepted in Soil Systems .

Author Response

We appreciate your collaboration in evaluating the article. Attached is a file with the justification/explanations for not including the analysis of the "region x land use x elevation" interaction. Please feel free to contact us with any further questions.

Author Response File: Author Response.pdf