General comment:

This manuscript addresses an important topic and provides a valuable comparison of soil inventory, RothC modelling, and eddy covariance approaches for estimating long-term SOC dynamics in cropland systems. The study is well designed and based on a unique long-term dataset.

The introduction is clear, well structured, and effectively motivates the study objectives. The methodology is described in a comprehensive and transparent manner, allowing the reader to fully understand the applied approaches and uncertainty analyses. The results are clearly presented and supported by informative figures and tables, while the discussion provides a balanced interpretation of the findings and places them appropriately within the context of previous research. Finally, the conclusions are well supported by the presented results and remain consistent with the scope of the study.

Overall, I consider this to be a good-quality manuscript that is suitable for publication after minor revision. The comments below are intended to further strengthen the presentation and interpretation of the results.

Comment 1: EC bias and energy balance non-closure

The manuscript presents a convincing comparison of SOC stock changes derived from soil inventories, RothC simulations, and EC-based carbon budgets, and clearly demonstrates a systematic divergence among these approaches. Furthermore, the discussion of the potential role of energy balance non-closure is thoughtful and appropriately cautious. The inclusion of both the carbon balance ratio (CBR) and energy balance ratio (EBR) provides useful support for the proposed explanation. As a possible avenue to further strengthen the discussion, the authors may consider briefly elaborating on whether the observed magnitude of EBR reduction could plausibly account for the discrepancy between EC-derived and inventory/model-based SOC estimates, even if such a relationship cannot currently be quantified with certainty.

Comment 2: Visibility of farmyard manure applications in Figure 4

Farmyard manure application is identified as one of the key drivers of SOC accumulation throughout the manuscript. In Figure 4, the caption indicates that vertical lines represent farmyard manure applications; however, these markers are difficult to identify or not readily visible in the current version of the figure. The figure could be improved by displaying manure application events more clearly (e.g., through stronger colours, thicker lines, symbols, or annotations). This would help readers visually relate management events to the temporal dynamics of NBP and SOC stock changes and would further support one of the central messages of the study.

Minor editorial comment:

Line 68 – CO2 – the lower index is missing

Line 106 – Please remove “2” form 20225

Line 128. – The dot is at a lower index; please change this format to normal.

Line 134 – Why is this hyphen used differently than the previous ones? Please unify it throughout the whole manuscript.

Line 228 - In my opinion, the paragraph should not begin with a mathematical formula. Please consider making this change throughout the whole manuscript to avoid this.

Line 379 – The version of R and the appropriate citation for this software should be provided.

General comment:

This manuscript addresses an important topic and provides a valuable comparison of soil inventory, RothC modelling, and eddy covariance approaches for estimating long-term SOC dynamics in cropland systems. The study is well designed and based on a unique long-term dataset.

The introduction is clear, well structured, and effectively motivates the study objectives. The methodology is described in a comprehensive and transparent manner, allowing the reader to fully understand the applied approaches and uncertainty analyses. The results are clearly presented and supported by informative figures and tables, while the discussion provides a balanced interpretation of the findings and places them appropriately within the context of previous research. Finally, the conclusions are well supported by the presented results and remain consistent with the scope of the study.

Overall, I consider this to be a good-quality manuscript that is suitable for publication after minor revision. The comments below are intended to further strengthen the presentation and interpretation of the results.

Comment 1: EC bias and energy balance non-closure

The manuscript presents a convincing comparison of SOC stock changes derived from soil inventories, RothC simulations, and EC-based carbon budgets, and clearly demonstrates a systematic divergence among these approaches. Furthermore, the discussion of the potential role of energy balance non-closure is thoughtful and appropriately cautious. The inclusion of both the carbon balance ratio (CBR) and energy balance ratio (EBR) provides useful support for the proposed explanation. As a possible avenue to further strengthen the discussion, the authors may consider briefly elaborating on whether the observed magnitude of EBR reduction could plausibly account for the discrepancy between EC-derived and inventory/model-based SOC estimates, even if such a relationship cannot currently be quantified with certainty.

Comment 2: Visibility of farmyard manure applications in Figure 4

Farmyard manure application is identified as one of the key drivers of SOC accumulation throughout the manuscript. In Figure 4, the caption indicates that vertical lines represent farmyard manure applications; however, these markers are difficult to identify or not readily visible in the current version of the figure. The figure could be improved by displaying manure application events more clearly (e.g., through stronger colours, thicker lines, symbols, or annotations). This would help readers visually relate management events to the temporal dynamics of NBP and SOC stock changes and would further support one of the central messages of the study.

Minor editorial comment:

Line 68 – CO2 – the lower index is missing

Line 106 – Please remove “2” form 20225

Line 128. – The dot is at a lower index; please change this format to normal.

Line 134 – Why is this hyphen used differently than the previous ones? Please unify it throughout the whole manuscript.

Line 228 - In my opinion, the paragraph should not begin with a mathematical formula. Please consider making this change throughout the whole manuscript to avoid this.

Line 379 – The version of R and the appropriate citation for this software should be provided.

This manuscript addresses an important question related to the assessment of long-term SOC dynamics in croplands by comparing soil inventories, RothC simulations, and eddy covariance measurements. The study benefits from a unique long-term dataset and combines several independent approaches to improve understanding of SOC changes under agricultural management.

The manuscript is generally well written, and the study design is sound. The methods are described in detail, the uncertainty analysis is comprehensive, and the results are presented clearly. The discussion is extensive and supported by relevant literature. The work has clear scientific value and will be of interest to researchers working on soil carbon monitoring, modelling, and greenhouse gas accounting.

Overall, I consider this manuscript suitable for publication after minor revision. The comments below are intended to help improve the interpretation of some results and strengthen the discussion.

Lines 57-58 The statement that NBP should theoretically correspond to changes in SOC stocks may be too simplistic. NBP and SOC stock change are related but not identical quantities, particularly in agricultural systems where carbon can be stored or removed through different pathways.

Lines 61-62: The uncertainty discussion focuses primarily on EC-derived estimates. Similar emphasis should be given to uncertainties associated with SOC models, including model structure, parameterization, and initialization.

Line 365: The justification for the  criterion needs more support. Why this threshold is chosen? Any sensitivity analyses has been applied?



Lines 355-363: Residues entering RothC are normally partitioned using a DPM/RPM ratio. For instance, for cereal straw it is 1.44 and for woodland litter is 0.25. The presented method focuses only on DPM decomposition. However, a substantial fraction of residue C enters RPM, especially for lignified residues. Consequently, the fraction of total residue carbon decomposed may be much lower than the calculated .



Line 377: "All monthly predictions are then pooled to derive typical statistical parameters." Suppose there are 10,000 simulations 240 months, Then pooling all values together creates  predictions. However uncertainty should generally be calculated for each month separately, rather than across all months.



Lines 370-379: It is assumed that . It was itself an assumption likely carrying uncertainty. This uncertainty should ideally also be propagated.



Lines 368-379: The Monte Carlo analysis appears to assume that all parameters and inputs are independent. However, this assumption may not always be realistic. For example, uncertainties in temperature and moisture modifiers may be related, decomposition rate constants may be correlated, and carbon inputs may be influenced by climate conditions. Ignoring these relationships can either increase or decrease the estimated uncertainty. The authors should acknowledge this assumption and discuss its potential effect on the results.

Lines 380-400: The comparison between NBP and ΔSOC RothC should be interpreted with caution because the two variables do not represent exactly the same carbon fluxes. NBP reflects the overall ecosystem carbon balance, including plant carbon uptake and respiration processes, whereas RothC only simulates changes in soil organic carbon based on estimated carbon inputs and decomposition. Therefore, agreement in source or sink classification may be expected, but close agreement in magnitude should not necessarily be assumed or interpreted as model accuracy.

Lines 598-610 : This section discussing subsoil carbon dynamics is informative and highlights an important limitation of the RothC simulations. However, the potential impact of subsoil processes remains largely qualitative. It would be good if the discussion could be strengthened by providing a simple estimate of the possible magnitude of subsoil carbon inputs based on the reported root biomass measurements. This would help readers assess whether subsoil dynamics are likely to explain a meaningful portion of the difference between NBP and ΔSOCRothC.

Finally, several parts of the discussion appear to compare NBP and ΔSOCRothC as if they represent the same quantity. However, NBP describes the overall ecosystem carbon balance, while RothC simulates changes in soil organic carbon stocks. It may be useful to remind readers that agreement in source or sink classification may be expected, but agreement in magnitude should not necessarily be expected. A short clarification in the discussion would help avoid potential over-interpretation of the comparison.