the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 24 Apr 2025
Improved management increases soil mineral-protected organic carbon storage via plant-microbial-nutrient mediation in semi-arid grasslands
Abstract. Soil organic carbon (SOC) storage in semi-arid grasslands is threatened by both climate change and land degradation, impacting food production and climate regulation. Improved management has been proposed to increase SOC stocks and overcome these challenges. However, the benefits of improved management practices in semi-arid regions are in question. Little is known about the effects of management on the functional components of SOC, particulate (POC) and mineral-associated organic carbon (MAOC), which are expected to respond differently, and about the pathways that mediate these responses, such as changes in vegetation and soil microbial communities.
This work analyses the effect of rotational grazing, legumes sowing and grazing exclusion on topsoil SOC, POC and MAOC stocks in Mediterranean wooded grasslands compared to continuous conventional grazing. Changes in plant diversity and morpho-chemical traits, soil fertility and microbial composition were also evaluated. A total of 188 plots were sampled in 9 farms across a wide environmental gradient.
More resource-acquisitive, nitrogen-rich and less lignified plant community, higher soil microbial biomass with lower Gram+/Gram- ratio, and higher soil fertility were associated with higher SOC storage, with similar impacts on POC and MAOC. Rotational grazing increased MAOC and total SOC stocks by 11 % compared to continuous grazing. This effect was mediated by an increase in soil fertility in the rotationally grazed paddocks. On the other hand, grazing exclusion reduced POC stocks by 12 % compared to continuous grazing. This depletion was mainly due to a reduction in microbial biomass and an increase in the C/N ratio of vegetation in non-grazed paddocks. Both POC and MAOC stocks tended to be lower at the warmer sites.
We conclude that rotational grazing can enhance long-term SOC storage in semi-arid grasslands, thereby increasing their resilience and climate mitigation capacity, whereas abandoning grazing could lead to SOC losses.
- Preprint
(2692 KB) - Metadata XML
-
Supplement
(1267 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-1711', Anonymous Referee #1, 05 May 2025
This manuscript presents an original and timely contribution to our understanding of soil carbon (SOC) dynamics in semi-arid grasslands, a topic of high relevance given the increasing threats of climate change and land degradation to these ecosystems. The study stands out for its comprehensive methodological approach, which combines field-scale experimentation with detailed plants, soil, and microbial analyses across an environmental gradient. The authors successfully integrate ecological, biochemical, and management dimensions to investigate how different grazing practices influence both particulate and mineral-associated organic carbon pools.
Despite the strengths outlined above, there are several methodological and interpretative aspects that require clarification or revision before the manuscript can be considered for publication.
Issues with chaOM fraction:
Figure 3a. The reported mean C/N ratio for the chaOM fraction is 4.41, and several values appear to fall between 0 and 2. These values are exceptionally low for any known pool of organic matter. I recommend that the authors clarify whether the C/N values presented for the soil fractions reflect total N or organic N (i.e., after subtracting inorganic N). If only total N was used, recalculating C/N based on organic N would likely provide a more accurate picture of the biochemical composition of these fractions.
The manuscript reports the use of a combined size-density fractionation protocol following Leuthold et al. (2024), in which the light fraction (<1.85 g cm⁻³) is separated from the heavier material. The latter is subsequently sieved at 53 μm to yield a coarse heavy-associated organic matter fraction (chaOM; >53 μm) and a fine mineral-associated organic matter fraction (MAOM; <53 μm). The chaOM and MAOM fractions were later combined and reported together as MAOM.
While the decision to include chaOM within MAOM appears justified based on the rationale provided in Leuthold et al. (2024), this approach diverges from the more widely accepted definition of MAOM as material denser than 1.6–1.85 g/cm³ and smaller than 50–63 μm (Lavallee et al., 2020). Consequently, this choice may limit the generalizability of the findings and their integration into broader syntheses. That said, since the chaOM fraction accounted for less than 5% of total SOM, the implications are likely minor.
Nonetheless, as emphasized by Lavallee et al. (2020), there is a pressing need for greater consistency in the operational definitions of POM and MAOM to facilitate comparability across studies. I recommend explicitly acknowledging that organic matter >53 μm is not typically included in the MAOM fraction.
L184-187. 2.2. Climatic variables. This section consists of a single sentence, which may not warrant a standalone subsection. I suggest either expanding it with additional context or incorporating it into a broader methodological section.
L248. SOC was already defined.
L249-250. It would be useful to specify whether inorganic carbon (e.g., carbonates) was measured or removed prior to C analysis.
Figure 4. Please add the slope of the regression line. This figure should be in the results section.
L299. This sentence should be included in another paragraph.
Figure 5
- The PCA plots refer to the components as "Dim", whereas the figure legend refers to them as "Axes". For consistency and clarity, I recommend using a single term.
- Axis 1 is labeled “C/N axis”. However, C/N is a single measured variable, and a PCA is not required to assess it alone. This axis also includes strong loadings from other traits such as ABG cellulose and CWM hemicellulose. Therefore, the axis appears to represent a broader concept, perhaps related to litter quality or decomposition potential, rather than just the C/N ratio. I suggest reconsidering the label to better reflect the multidimensional nature of the trait composition.
- A similar issue applies to Axis 3, labeled “Lignin axis”. Lignin is again only one of several traits contributing to the axis. Since this component also captures variation in ABG hemicellulose and CWM LNC, among others, it likely reflects a broader litter chemistry or quality gradient rather than lignin content per se.
- In panel (h), ABG hemicellulose is more negatively correlated with the axis than CWM SLA, yet only CWM SLA is shown in the schematic.
- Finally, I recommend reducing the number of figure panels by integrating each schematic (panels e–h) below the corresponding PCA biplot (a–d). This would streamline the presentation while preserving the valuable visual summary of trait syndromes.
L322. … bulk density (BD). What was the range of BD? Was it similar across treatments? If not, equivalent mass corrections should be considered.
L340. Could you please clarify what is a direct and indirect effect?
L364-365. “…informing on lignin content and vegetation productivity, was negatively correlated with the POC, MAOC and SOC stocks and the microbial biomass”. Lignin content was negatively correlated with vegetation productivity. Therefore, those variables were negatively correlated with lignin, but positive correlated with vegetation productivity.
L430-432. “The silvopastoral character of our farms could explain the low MAOC/SOC ratios, as litter from scattered trees increases carbon stocks in woody grasslands, especially in the POM fraction”. This explanation is plausible; however, your results show that lignin and C/N were negatively correlated with POC. How do you reconcile this apparent inconsistency?
L439. I don’t see a saturation curve in Fig. 3 b. Figure 4 shows a better way to illustrate the C-saturation deficit. It would be interesting to compare the results with the maximum C-loading found by Georgiou (~ 86 mg C g silt+clay).
L497-498. However, roots promote MAOM formation not from root biomass but from rhizodeposition (see Sokol, N.W., Bradford, M.A., 2019. Microbial formation of stable soil carbon is more efficient from belowground than aboveground input. Nature Geosci 12, 46–53. https://doi.org/10.1038/s41561-018-0258-6; Villarino, S.H., Pinto, P., Jackson, R.B., Piñeiro, G., 2021. Plant rhizodeposition: A key factor for soil organic matter formation in stable fractions. Science Advances 7, eabd3176.)
L546. In line with my previous comment, bulk density values and changes should be reported. Previously, the acronym BD was used for apparent density.
L590. Mineral-associated organic carbon was previously defined as MAOM.
Citation: https://doi.org/10.5194/egusphere-2025-1711-RC1 -
AC3: 'Reply on RC1', Victor Rolo, 16 Jun 2025
Firstly, we would like to thank you for the time and effort you have put into this review. We believe that all your comments have been very valuable in improving the first manuscript submission.
In the attached document, we have included your comments in black and our response to each comment immediately below in blue color.
-
RC2: 'Comment on egusphere-2025-1711', Anonymous Referee #2, 15 May 2025
This is a well written manuscript with a good concise structure, useful figures and relevant results for the soil and grassland science communities. The authors have performed an wide study on how grazing intensity and management affect soil organic C stock (total C and MAOC and POC fractions) in semi-arid woody grasslands across 3 different climate regions in Spain. Further, the study analyses both direct and indirect mechanisms that are driving the observed changes, combining several measured variables in a structural equation model. This approach is a strength of the manuscript, and together with the specific findings it brings new insights too the scientific field that are also relevant for policy makers and farmers. Thus, I find the paper relevant for publishing in SOIL after some requests and suggestions have been clarified, specified below. My main concern is the measuring depth of 8 cm, which limits certain conclusions. This limitation should be discussed more clearly than what has been done in the present version of the manuscript, and the reasoning behind choosing this depth should be made clear.
Specific comments
Abstract:
Please describe the reason for only sampling to 8 cm depth in the introduction and/or M&M, and consider mentioning the depth in the abstract or, alternatively, write “upper topsoil” instead of just “topsoil” in the abstract. I believe there might be arguments for choosing a relatively shallow depth, but the arguments are not mentioned nor discussed (microbial dynamics predominantly occurring in the upper topsoil? No/low tillage, etc.). In terms of the management effects on soil C stocks, the conclusions of the study are limited to the upper 8 cm - which should be discussed.
Keywords/introduction:
I find the expression “Legume enrichment” slightly unclear/confusing. In my opinion, it connotes with “isotope enrichment experiments”, where the legumes themselves are isotope enriched. Consider if writing “Legume sowing” og “legume inclusion/addition” as keyword and where it is mentioned in the manuscript (e.g., line 58), could be an alternative.
L109-112. Species richness effect on MAOC/POC fractions in managed grasslands have been investigated recently by Mortensen E.Ø., Abalos D., Engedal T., Lægsgaard A.K., Enggrob K., Mueller C.W., Rasmussen J. (2025) Smart mixture design can steer the fate of root derived carbon into mineral-associated and particulate organic matter in intensively managed grasslands. Global Change Biology 31: e70117. DOI: 10.1111/gcb.70117. The study observed no effect of species richness per se, but effects of plant functional group (grasses vs. legumes). Please consider this in the introduction and/or discussion section, i.e., how that compares with the findings of your study.
Material and method:
L162: What is the lower range of time for the legume enrichment/sowing? Does the study also include paddocks where legumes have been sown within the last year or 2, or is there a minimum of “incubation time” for the effects to be observed? E.g. “5-2 years”?
L163-165: It would be good with a few more words on the legume species that are chosen in those mixtures that are seeding on the farms, instead of only mentioning the broad genus names and referring to another study. Is it always the same seed mixtures, and/or on which criteria are the seeded species selected? Which are the most dominating legumes in the farms that are already established before legume sowing (based on the botanical assessment)? Here, I am thinking specifically on the root systems, whether shallow and/or deep-rooted legumes existed before/are sown. I think this is relevant for the reader, especially because of the no/negative effects of recent legume sowing on SOC and MAOC stocks. The type/species of sown legumes may have an influence on which effects legume sowing induce/not induce in the upper 8 cm of the topsoil – and which potential effects below 8 cm may not be covered in this study. This caveat should be mentioned when discussing the results.
L202: I don’t find any results on the aboveground species botanical composition. In line with the comment above, a (supplementary) table with species composition, or at least mentioning the dominant species, would increase our understanding of the system. Does this vary too much between the farms to provide this information, or are there some general species (non-legumes and legumes) that dominate/characterize the farms used in the study?
L231: As mentioned above, 8 cm is quite shallow when considering the effects of species on soil C. Please, argue why 8 cm depth was chosen instead of e.g., 15, 20, or 25 cm that would include the effect on a larger part or the entire topsoil layer.
L281: Is the reference (“as shown in eq. 1”) referring to the same equation that are used to calculate OC in soil C fractions? If so, please make this clear, as it does not seem obvious that microbial biomass stocks can be calculated by PLFA via the same equation.
L304: Does “the latter being considered a good proxy for the lignin content (Van Soest et al., 1991).” refers to ADL – or to “ADF minus ADL”? If it’s the first, I suggest clarifying it by writing: “… ADF minus ADL, ADL being considered a good proxy for the lignin content (Van Soest et al., 1991)”.
Results:
L358: Please check if the mentioned negative correlation between the Fungi/Bacteria ratio and MOAC stock was statistically significant? It does not look like that in the Figure 7b annotation, and in that case, you could argue for a tendency while just stating that they are negatively correlated is wrong if it is not statistically significant. Same comment for line 455 where this correlation is mentioned again.
L358 and line 359: For both sentences here, you would ease the readability by adding an explanatory sentence in both places, e.g. such as “… meaning that a higher proportion of fungi over bacteria correlated with higher MAOC stock” – or what direction the correlation may have. Especially when ratios are negatively correlated, it becomes a lot of twisting to interpret the direction of the effects, unless you are very familiar with the parameters.
Discussion:
L417: I recommend writing something like “… long-term carbon storage in the upper soil layer/topsoil”, or in other ways mentioned the depth restriction of the conclusions already here.
L437-443: This is an important point and paragraph! Good.
L477-478 and 484-486: These are two places where it may be relevant to discuss whether your findings align with Mortensen et al., 2025, mentioned above (doi.org/10.1111/gcb.70117).
L541-551: In this section (the effects of legume sowing), it is particularly relevant to discuss the shallow sampling depth of 8 cm, and making us aware of the limitations of the conclusions. Sowing new species will most likely also affect carbon and nitrogen dynamics below 8 cm, and thus the total SOC/POC/MAOC stocks of the soil profile. This aspect could also be discussed in a separate paragraph for all management effects, but at for the effect of legume sowing.
L593: Please change from “are” to “were” (saying that your study found this - not common knowledge), or if generalizing this statement to more than your study, at least define the limit of this conclusion to the upper topsoil layer and to semi-arid grazed grasslands. As it stands now it is a very strong generalization, which may not necessarily be the same in other regions, agricultural systems, nor if a deeper soil profile was investigated.
Wording and typos:
L97: Typo. Change from “MENS” to “MEMS”
L119: Consider writing “… biochemical and morphological traits”
L213: I assume the last word of the sentence should be in past tense: “…measured…”
L302: Change to “These analyses…” (plural)
L322: Check grammar. I would write “Nutrient concentrations in each soil were…”
L330: Insert comma: “SOC, POC and MAOC”
L372: I assume this should be in past tense to align with the rest of the results (“increased”) not indicating general knowledge which would be in present tense.
L385: Delete “it”, writing just “As expected, being negatively ….”
L559: Change to “Our results are in line with…”
L617-621: Please check the use of “” signs. There seem to be lacking some "" around some parts. Also align first names in this section to be either full first names or just the first letter.
Figures:
Fig. 2: In the figure caption, do you mean to write “References embedded in the figure or Fig. 2….?”
Fig. 4: There might be missing a word in the end of the first line: “… in the studies ___?___ soils and… ”. Or split the sentence into two. It does not make sense how it is currently reading.
Fig. 5: Please increase the font size of most elements in the figure. As it is now it’s not possible to read all variable names in the PCA’s in a 1:1 printout, tricky to read the legends, and very tricky to read the blue/red text in the gradient explanations/axis explanation (in the lower part of the figure). It is a shame since the figure contains a lot of good information. Although it is nice that the text coloring follows the legend/gradient in the PCA’s, it also makes it harder to read the text. Maybe the coloring can be kept if the text size increases – or vice versa if text is darker.
Fig. 6: Axes font size within legends is very small, making it impossible to read certain words in a 1:1 size. Also, consider changing “R2” to R2” in the center circle (stock changes). Other than that, it is a key useful figure. In the figure text, you can delete one “s”, thus writing “…by factor type…”.
Supplementary:
Table 1: Please include management acronyms in this table, as done in figure S3.
Figure S1: As for figure 5, please increase font size. Font color could be OK if size increases.
Figure S3: Increasing font size of labels would make it possible to read them.
Figure S5: Please provide the unit for the content in the figure caption - or if it is centered and scaled / indexed, mention this in the figure caption as well.
Citation: https://doi.org/10.5194/egusphere-2025-1711-RC2 -
AC1: 'Reply on RC2', Victor Rolo, 16 Jun 2025
Firstly, we would like to thank you for the time and effort you have put into this review. We believe that all your comments have been very valuable in improving our manuscript.
In attached document, we have included your comments in black and our response to each comment immediately below in blue color.
Please notice that, motivated by reviewer 1's justified concerns about the data for the chaOM fraction, we have reanalysed 111 samples from that fraction, which has slightly changed some of the study's results. These changes are detailed in the response to reviewer
-
AC2: 'Reply on RC2', Victor Rolo, 16 Jun 2025
Firstly, we would like to thank you for the time and effort you have put into this review. We believe that all your comments have been very valuable in improving our manuscript.
In attached document, we have included your comments in black and our response to each comment immediately below in blue color.
Please notice that, motivated by reviewer 1's justified concerns about the data for the chaOM fraction, we have reanalysed 111 samples from that fraction, which has slightly changed some of the study's results. These changes are detailed in the response to reviewer
-
AC1: 'Reply on RC2', Victor Rolo, 16 Jun 2025
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 515 | 141 | 20 | 676 | 27 | 20 | 33 |
- HTML: 515
- PDF: 141
- XML: 20
- Total: 676
- Supplement: 27
- BibTeX: 20
- EndNote: 33
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1