Maier et al studied the distribution and mechanisms controlling SOC distribution along a geochemical gradient of soil parent materials. The manuscript is overall well-written and the analyses are appropriated. The study compared plant inputs, soil characteristics and parent material in explaining the SOC stocks along the geochemical gradient and pointed out that “Parent material geochemistry – and not plant input - as the primary element shaping soil organic carbon stocks in European alpine grasslands”. Please find below my major and minor concerns.

1. From the experimental results to European alpine grasslands. Because the experimental setting is along the geochemical gradients of soil parent materials, so the sites are selected to represent the difference in parent material. Comparisons among selected sites showed that parent material is important. Does it fair to conclude that parent material is more important than input for (entire) European alpine grasslands? Would the conclusion be different if you select the experiment sites across a plant input gradience?
2. Soil fertility vs. soil mineralogy model. The authors build random forest models with 42 samples and multiple explainable variables. Do these models face overfitting issues? What insights can we gain from these analyses, especially differentiating soil fertility vs. mineralogy? I feel the rationale is not very well justified. Why not have plant input as an explainable variable, and why not plant input, fertility and minerology together explain SOC variations?
3. Soil aggregates. The study found that microaggregate contributes to a large (>50%) portion of bulk SOC, and inferred those sites with metal oxides favored the formation of microaggregate. Are there approaches to provide more direct evidence from this experiment?
4. Please double check on figures. There seem swaps. For example, Figure 3a, Gneiss, the crosshatch and solid part seem did not follow the texts. The crosshatch is much big, but in the text, it states that the wood (solid) is bigger. Figure 5 c, should it be the soil fertility model, but in the captions, it states “soil mineralogy model”. This swap also affects Figure 5d   

Maier et al., 2025: Parent material geochemistry – and not plant input - as the primary element shaping soil organic carbon stocks in European alpine grasslands

General overview:

Maier et al., 2025 compare observations from European alpine grassland soils developed in five geochemically-distinct parent materials to assess how plant biomass, soil fertility, or geochemical variables are related to SOC stocks. The authors then use physical fractionation and ¹⁴C analyses to evaluate the distribution of carbon and its persistence (in the bulk soil). The study reconfirms a well-established correlation between Fe, Al, and Mn concentrations, extracted with classical-sequential extractions, and SOC stocks, while also reporting a large contribution of the microaggregate fraction to bulk SOC. Overall, I enjoyed reviewing this article and found that their data was well presented, and the study, well referenced. I thought that the study benefited from a well-thought out sampling strategy, which took into account landscape heterogeneity within their sites, and the difficulties of getting representative samples from grasslands in environments with complex topography.  

My comments regarding the manuscript are largely minor, but are numerous. Instead of separating my comments into major or minor revisions, I’ve instead presented the comments as a line-by-line evaluation, with more general comments at the beginning of each section. My main two general comments are in connection to the language and figures. Please carefully evaluate your terminology throughout and be more specific with its usage in the manuscript. I’ve also made some suggestions on your existing figures and would ask for the inclusion of more supplementary figures to support your narrative and discussion. With these corrections, this paper will shortly be ready for publication in Biogeosciences.

Title:

• Is element the right word here? Would driver, (controlling) factor, or variable be better suited?

Abstract:

• Line 4: Is distribution the right word here? Or should it instead be SOC stocks? And is it SOC stabilisation mechanisms or instead geochemical variables?
• Line 10: At the moment, this reads that soil fertility governs plant C inputs and soil mineralogical characteristics. I’d suggest instead combining “soil fertility and soil mineralogical characteristics, which govern plant C inputs and control C stabilisation respectively, play equally critical roles”….
• Line 11: Could the F¹⁴C values be mentioned earlier? As the previous sentence quite nicely summaries the paper.

Introduction:

• Line 19: Is this in reference to a specific study? It seems so. If so, please add the citation.
• Line 24: This sentence has an abrupt break after understudied, “remain understudied” would probably be better suited at the end of the sentence.
• Line 32: and short-range order minerals / oxy(hydro)oxides?
• Line 40: As a suggestion it could be nice to clarify and conclude why this paragraph has been important to your narrative in your own words. It ends rather abruptly. This occurs throughout the manuscript.
• Line 41-49: Climate change could also enhance the decomposition of C at elevation and it would be good to present this critical counterpoint. By increasing the temperature of soils, increasing the efficiency of decomposers, their extra-cellular enzymes, changing moisture regimes, and increasing the mineralisation / respiration of SOC. It would be good to provide both sides of the argument here. I believe that Garcia Franco et al., (2024; Geoderma 442, 116807) may be a good reference in this case.
• Line 51: I’d suggest a paragraph break here to separate your study design from the knowledge gap and then combine the latter part of this paragraph with the following paragraph containing your hypotheses.
• Line 60: I’d also suggest being more succinct in your hypotheses in this final paragraph of the introduction.
• Line 70:  conjoined parentheses - ) ( could be combined with a semicolon throughout. (POM; Kleber et al., 2015; Torn et al., 2013). If using Endnote, this can be achieved with the prefix function. Please change this throughout.

Materials and Methods:

• Line 82: Nested parentheses - Brackets in brackets are typically square brackets or a different format. Please change this throughout.
• Line 89: This is an open question, but does this study assume similar age of each soil type since deglaciation? How could slight variations in your soil forming factors have influenced your interpretations? Is it really geochemistry that is defining your SOC stocks or just time and the expression of pedogenesis?  

I see that aspect and slope angle was kept as constant as possible with your study design, but could you please add it to Table 1 if you have it?

Furthermore, I will trust your classification, but is this definitely a ~15 cm thick Oh horizon atop a R horizon? There is no Ah horizon here?

Principal qualifiers are capitalised.

• Line 90: Calcareous?
• Line 94: If I’m not mistaken, Calcaric is the principal qualifier in connection with Cambisols, again capitalised. Calcic would make your profile a Calcisol (Calcic horizon) and is not typically applied to a Cambisol.
• Line 99: September,
• Line 101: You can remove the % after 40 as it is repeated after 60.

Trough is a strange old English word in this context, meaning a long narrow open container for animals to consume from, or a channel for moving water. I don't think that you referring to tillage (trough) here? Is it instead micro-elevational changes at your grassland? Please clarify.

• Line 117: Ground would also be applicable or just milled.
• Line 126-128: I might have missed this, but I didn’t see mention of this in 2.3.8. What was the effect of their inclusion?
• Line 142: Have you got a ref for the pH measurement? It’s normally achieved with different timing and soil-to-solution ratios so it would be good to include the ref.
• Line 143: Why were Si, Ti, and Zr measured with XRF and the rest of the elements measured with a pseudo-total digest using aqua regia? Didn’t you get these elements from the XRF measurements?
• Line 154: It seems like only the Cw horizon of the Marl soil had a pH higher than 6. Did you also treat the Oh horizon of the Dolomitic Leptosol? Did you measure the stable isotopic C composition also? Your work assumes that carbonate removal was complete, but was it?
• Line 176: This first sentence needs to be corrected. “To identify reactive metal phases that can interact soil C…” Be careful with the differentiation of correlation and causation throughout, particularly when referring to C stability.
• Line 180: I appreciated your disclaimer here regarding the usage of sequential extraction and the table below. Even though these are well established extractions, it would however have been good to include how they were extracted, similar to your description of the pseudo-total digests. The CEC extraction could also be more detailed. 1 g of soil was extracted in 100 mL…

Were the ICP-OES measurements made with an internal standard? What was the relative standard deviation? Did you measure duplicates? How reproducible were your observations? 

• Line 187: I’d suggest splitting the calculation of the fertility indices section from the F¹⁴C measurements as they’re largely unrelated.
• Line 201: ca. is typically italicised like vs., throughout.
• Line 206: Soil persistence?

Section 2.3.8:

• How did you test for normality?
• Did you evaluate changes with depth also? And if so, how do your statistics account for the spatial autocorrelation with depth? Could linear mixed models have been used to evaluate changes with depth?
• A principal component analysis could also be informative to demonstrate the differences between your soil types / geochemical gradient and its relationship to your various bulk measurements, even if it’s only in the supplementary information. This could also provide interesting insight into how your geochemical variables are associated with your C variables. Is this coming in a separate manuscript? If not, I’d suggest you include this and more figures in the supplementary information, exploring depth gradients and the distribution of your measured variables.
• SOC prediction models - what was the effect of the exclusion of clay content from your prediction model? It seems strange to include it in both the soil mineralogy and soil fertility models if it has such a strong explanatory power. It seems more relevant to the soil mineralogy prediction model. Please test this or provide a strong rationale for its inclusion in both models.
• Are your averages presented as ± standard error. It’s not stated explicitly, except in captions from what I could see. Maybe I missed it, but then the point is it could be made clearer.

Results:

Be careful with reporting discussion in your (separated) results section. As you have separated the sections, please keep references and interpretation in the discussion. See line 274:279 for an example.

• Line 261: Are widest and narrowest the correct terminology here? I would suggest not at is refers to a range and it’s counter to your observations and their variability, reported just after, as the Marl site has a higher variability. I believe it should be a higher or lower C:N ratio.
• Line 269: Superscript ^-1.
• Line 274: this is discussion as per my above major point and should be removed from the results section.
• Here I believe you are referring to better conditions for plant growth or something specific. What are ideal conditions? They are subjective.
• Line312: as the oldest values, do you believe there was any geogenic influence on the F¹⁴C values of the Cw horizon at your Marl site or are you certain that decarbonation was complete? Also your Oh horizon in the dolomite soil has an older F¹⁴C value than Ah horizons from other soil types? Could you please discuss this.
• Line 319:324: I was surprised at how much of your bulk SOC was represented in the microaggregate fraction in these soil types. I guess it is inline with other observations from Chilean alpine grasslands using a similar fractionation scheme. It’s also interesting that it remained relatively constant with depth relative to your two other fractions. This warrants more detailed discussion later in the manuscript.
• Line 319: Is it an importance of a fraction to bulk SOC? Or the distribution of C amongst fractions (g g^-1  or g 100 g^-1).
• Line 324: but increasing with depth.
• Line 328: In your methods you refer specifically to prediction of SOC stocks, here you predict SOC content. Please be specific and consistent with your terminology throughout.
• Line 329: R² typically.
• Line 332: Again, I’m not sure that you should include clay in both models, particularly if it’s such an important predictor. I would suggest that you test its exclusion and then if you choose to include it, defend its inclusion in both models more thoroughly. 

Discussion:

As a general comment, it would be good to point the reader to any existing literature or data on these specific sites with a paragraph in the discussion, discussing the observations that have been made and how they compare with your data, to give the reader better context. It would also be informative to compare the results of authors who have used a similar fractionation scheme (building on your comments about Wasner et al., 2024) and compare the results of your fractionation, the distribution of SOC amongst fractions in your system with theirs, and discuss how differences in fractionation scheme or soil systems could have driven these observations.

• Line 371:373: This is surprising, I wonder how much of it is related to the fact that shrubs are pioneer species that prefer sandier soils, and produce more biomass in your system, relative to a stabilising influence of sand on high OC stocks (black sand soils being a notable exception).
• Line 375: This could be a good break of a paragraph rather than Thus, and Therefore, in subsequent sentences.
• Line 378: How much of this is driven by the soil thickness and stage of subsoil development, relative to a purely geochemical influence.
• Line 381: Relatively speaking, dolomite is less soluble than calcite, another important mineral in alpine grasslands in the Alps.
• Line 389: Does the combination of PP and AO not contradict the findings in Line 335 - that the poorly crystalline AO oxides correlated negatively with SOC, while PP positively correlated with SOC prediction in your model?

I would suggest including more figures, at least as supplementary figures to help display the distribution of your bulk geochemical data and its correlative relationship with SOC, it may help the reader understand the distribution of your data.

•  Line 412: I hope that the authors have collected F¹⁴C data on these fractions as it would be great to see this in a subsequent manuscript.

I wonder again if this could be related to an incomplete carbonate removal in the Marl site.

• Line 422: I think that this is “rather stabilization-[driven] than input-driven, adding the hyphen creates a connection to -driven, otherwise it reads slightly bizarrely.
• Line 439: I don’t believe this agrees with your abstract? Please be consistent with the use of numbers and terms throughout.
• Line 446: It seems that this is a rather important part of your discussion that stops abruptly. I would suggest building on this and exploring its connection not only to Wasner et al., who used a similar fractionation scheme, but also other studies that have used different fractionation schemes to explore its relevance and reasons. Every fractionation scheme results in operationally defined pools, but it would be good to discuss the relevance of this and compare it to other studies and systems.

Figures & Tables:

General: Some tables are included in the text and others not, I suspect this is just a formatting issue in early submission. Table 4 pops up earlier in the text then the other tables, and the table numbers should instead increase sequentially. Unless this data is being explored in a subsequent manuscript or is already under review, please consider adding more supplementary figures to explore your bulk geochemistry dataset.

• Fig. 1: I would suggest dropping the colour in the map background to more clearly highlight where the samples come from. A grey scale for the coarse hillshade/digital elevation model in the background would highlight the site colours more clearly.
• Fig. 2: As a suggestion, rather than reporting the rooting zone, do you think it would be interesting to present root coverage as a relative percentage of your profile surface with depth? Don’t let this suggestion crowd your figure unnecessarily, but it could be informative.
• Table 1: I’d add slope aspect and character to this table. Even if they were controlled for, it would be nice to present them and how this could have influenced your interpretations.
• Fig. B1 & B2, can you please remove the grid lines from the plots in the upper right so it’s easier to read the parent material type and numbers.