This manuscript presents a well-done designed and highly relevant study that investigates the effects of intensified freeze-thaw cycles (FTC) on winter nitrogen dynamics in temperate grasslands. The application of an in situ ¹⁵N tracer approach across two contrasting grassland sites represents a significant methodological strength, providing direct insights into the fate of winter N sources. The central finding—that intensified FTC restructures winter N availability by enhancing microbial retention and altering plant competitive hierarchies, rather than causing simple N losses—is novel, compelling, and has important implications for predicting ecosystem responses to winter climate change. The study is timely, addresses a critical knowledge gap, and possesses innovation. 

While the scientific foundation of the work is strong and the conclusions are broadly supported by the data, the manuscript in its current form requires significant revision to fully realize its potential. I am enthusiastic about the potential of this manuscript to make a valuable contribution to the field. The necessary revisions are needed and primarily focused on presentation and interpretation. I am confident that after a thorough revision addressing the points above and those detailed in the specific comments, this manuscript will be suitable for publication in BG journal.

Major concerns:

(1) Structural and Narrative Flow: The organization of the Results and Discussion sections could be optimized to create a more logical. Specifically, the order of presenting findings could be rearranged to better guide the reader from the ecosystem-level outcome (N retention) down to the underlying mechanisms (soil processes, microbial uptake, plant competition).

(2) The Results section currently contains lengthy lists of percentage changes.

(3) While generally clear, the manuscript requires a thorough proofread for grammatical consistency, conciseness, and precise scientific terminology.

(4) Discussion: The Discussion would benefit from stronger integration between sections (e.g., explicitly linking microbial decoupling to plant responses) and a more focused interpretation of the statistical analyses (e.g., highlighting the key drivers from the random forest analysis, rather than listing correlations). 

Specific comments:

Introduction:

This Introduction presents a solid foundation with a clear research gap and well-defined hypotheses. The following suggestions aim to enhance the logical flow, sharpen the focus on the core scientific problem, and strengthen the overall narrative leading to the study’s objectives.

The transition between some paragraphs could be smoother to create a more compelling "storyline." For instance, the jump from the importance of winter processes (Paragraph 1) to global warming (Paragraph 2) could be more seamlessly connected.

Paragraph 3 effectively lists the competing processes but could be restructured to more sharply emphasize the central paradox or tension.

The important fact that temperate grasslands "cover nearly 40 % of China’s terrestrial ecosystems" is currently placed after the hypotheses, where it serves less of a motivational purpose.

The hypotheses are well-founded but contain extensive lists of referenced mechanisms, which can make them appear somewhat list-like and dilute their core predictive statement.

M&M:

The rationale for the FTC treatment levels (+6 and +12 cycles) is somewhat scattered between the site description and the experimental design sections. The reader must combine information from Table 1 and the text to fully understand the basis for these choices.

The sampling description in section 2.3 contains some ambiguity. It mentions excavating soil blocks "containing dominant plant species" (suggesting targeted sampling) but also collecting soil cores "randomly" (suggesting random sampling). It is unclear if plant and soil samples were co-located within the same 1m x 1m subplot.

Section 2.1 directs the reader to Table 1 but then proceeds to list much of the same data (precipitation, temperature, plant species) in the text. This creates redundancy.

Results:

Subsections like 3.2 and 3.4 contain long lists of percentage changes (e.g., "increased by 25.0% and 24.0%... increased by 44.3% and 58.6%...") without high-level synthesis.

Section 3.5 reads like a list of correlation coefficients, which is difficult to interpret.

Discussion:

The last paragraph of 4.4 contains a slightly repetitive list of reasons for the subordinate species' disadvantage. The powerful concept of "temporal niche partitioning" is present but could be emphasized as the overarching explanation.

Figure 7:

Please remove the plus signs (+) preceding the percentage values on the y-axis labels .

Minor comments:

1. Abstract Line 34: Modify to read “employing the dual-labeled isotope”.
2. Introduction Line 37: Change the period (.) before “while” to a comma (,).
3. Introduction Line 57: Modify to read “FTC results in distinct competitive environments”.
4. Line 58: Change “present” to “exhibit”.
5. Line 59: Change “,” to “-”.
6. Line 82: Modify to read “(1) intensified FTCs reduce winter N retention through three primary mechanisms: (a) physical disruption of soil aggregates that enhances N mobility, (b) root damage that impairs plant N uptake capacity, and (c) microbial cell lysis that leads to N leaching and denitrification losses;”.
7. Line 82: Modify to read “intensified FTC would cause differential utilization of winter N sources among plant species. This effect is mediated by interspecific differences in three key traits: competitive ability, root system architecture (particularly rooting depth and winter root activity), and growth phenology (temporal niche partitioning).”
8. Line 112: Modify to read “During the study period, the meadow steppe had a persistent snow cover that reached a depth of 20-25 cm in late winter”.
9. Line 114: Replace “exhibited” by “had”.
10. Line 117: Replace “underwent” by “experienced”.
11. Line 121: Modify to read “This contrast enables a comprehensive assessment of...”.
12. Line 125: Modify to read “...the predominant soil type is loam in the meadow steppe and sandy loam in the sandy steppe...”.
13. Line 132: Replace “(Trin).” by “(Trin.)” and replace “(L).” by “(L.)”.
14. Line 133: It seems that there is extra space between the figure (70) and the unit (%).
15. Line 155: Is it wrong here? If the amount of solution added is 120 mg ¹⁵N m^-2, then it should be 600 mg ¹⁵N L^-1.
16. Line 226: Replace “contrast” by “difference”.
17. Line 227: Replace “are” by “were”.
18. Line 231: Supplement the molar concentration of the K₂SO₄solution, modify to read “...by shaking with 60 mL of a figure M K₂SO₄..”.
19. Line 233: Replace “elemental analyzer” by “total organic carbon analyzer”.
20. Line 234: The conversion coefficient is abrupt, and it is integrated with line 225 to change: Microbial biomass C and N were calculated by dividing the differences in extractable C and N between fumigated and non-fumigated samples by a conversion factor of 0.45.
21. Line 264: R package “rcorr” should be combined with version information. It is recommended to read: “Spearman correlation coefficients between variables were calculated using the rcorr function (in the Hmisc R package).”.
22. Line 268: Replace “above mentioned” by “above-mentioned” and replace “analysis” by “analyses”.
23. Line 271: Replace “SigmaPlot, 14.0” by “SigmaPlot 14.0” and replace “R Studio” by “RStudio”.
24. Line 277: Replace “lowest” by “minimum”.
25. Line 297: Modify to read “In contrast, neither LFTC nor HFTC significantly affected NO₃⁻-N concentrations or net nitrification rates at either site during the study...”.
26. Line 302: Replace “LFTC” by “HFTC”.
27. Line 311: Modify to read “In contrast to the significant effects of HFTC, intensified LFTC had no significant impact on the shoot or root biomass N of the selected plant species at either site (Fig. 5a-f).”.
28. Line 315: The result section can be more concise and incorporate duplicate structures. Modify to read “HFTC significantly reduced biomass N in the perennial rhizome grass...(shoot: ...; root: ....)...” The sandy grassland is also changed in this way.
29. Line 325: Replace “was highest” by “peaked”.

30. Line 338: Modify to read “In contrast to the positive effects on microbial recovery, HFTC significantly reduced plant ¹⁵N acquisition in both grasslands. LFTC had no significant effect on plant ¹⁵N recovery ...”.

31. Line 353: This paragraph also simplifies the language.
32. Line 363: This section can be more concise and incorporate duplicate structures.
33. Line 364: Modify to read “In the meadow steppe”.
34. Line 367: Is it net nitrification rate ornet nitrification rates? The whole article should be unified.
35. Line 369: Modify to read“plant ¹⁵N acquisition showed the strongest positive correlation with soil temperature, while bacterial biomass, MBC, soil DOC, soil moisture, soil NO₃^--N levels, and soil total C also exhibited significant positive correlations.”.
36. Line 370: Replace “dissolved organic C” by “DOC” and replace “microbial biomass C” by “MBC”.
37. Line 372: Modify to read“…soil total N content and net ammonification rate displayed...”.
38. Line 375: Modify to read“In the sandy steppe”.
39. Line 375: Accordingto correlation, rank the indicators from largest to smallest and long indicators are replaced with abbreviations.
40. Line 377: It seems that a space is lacking between the “C” and the “and”.
41. Line 418: Replace “with”by “as”.
42. Line 421: Replace “.”by “,”.
43. Line 440: Replace “luxuriously utilize”by “engage in the luxury consumption of”.
44. Line 446: Replace “triggers shift”by “triggers a shift”.

 General Comments

This manuscript addresses an important topic in ecosystem nitrogen (N) cycling under changing winter conditions, using a ¹⁵N tracer approach to examine the fate of added nitrogen in plant, microbial, and soil pools following intensified freeze-thaw cycles (FTCs). The experimental design and tracer methodology are robust, and the study offers valuable insights into seasonal N allocation across ecosystem components.

However, several broader ecological interpretations—particularly those concerning ecosystem-level N retention, plant-microbial interactions, and species trait-based responses—go beyond what is directly supported by the data. Key processes such as N leaching, gaseous emissions, and root damage are not measured, limiting the conclusions that can be drawn about system-level outcomes and mechanisms. Additionally, there are some inconsistencies between the hypotheses, measurements, and interpretations that should be addressed.

Overall, the manuscript would benefit from a clearer distinction between the results directly observed in the data and the mechanisms proposed to explain them. The interpretation of the findings should be more cautious, and the methodological limitations of the study should be more explicitly acknowledged. Major revisions are needed to ensure that the hypotheses and conclusions accurately reflect the data obtained from the experiment, as some interpretations currently extend beyond what the results support. With these changes, the study could make a valuable contribution to understanding nitrogen cycling in seasonally frozen grassland systems.

Specific Comments

1. Elongation of FTC:

The manuscript assumes that climate change will lead to an elongation of FTCs. However, it is not clear why elongation rather than an earlier onset or other changes in FTC dynamics is expected, especially in temperate grasslands in China. We suggest the authors clarify this assumption and provide relevant references supporting the prediction of FTC elongation in their study region. This will strengthen the rationale for the study design and its climate relevance.

2. Hypothesis (1):

The first hypothesis posits that intensified FTC would reduce retention of winter N resources due to physical disruption of soil aggregates, root damage impairing plant uptake, microbial cell lysis, and subsequent N leaching and denitrification losses. However, the study does not include direct measurements or assessments of the mentioned parameters, so it is not possible to robustly test this hypothesis. We recommend revising the hypotheses to focus on mechanisms and processes that are directly measured or can be reasonably inferred from the data.

3. Losses of winter N sources:

The use of ¹⁵N tracers allows tracking of the fate of added labeled nitrogen, but it does not account for the dynamics of native, unlabeled N pools that may be mobilized during freeze-thaw cycles (e.g., through microbial lysis or mineralization of soil organic matter). This is particularly important given the observed increase in soil NH₄⁺ concentrations, which likely reflects the release of native N rather than enhanced retention of applied ¹⁵N. The conclusion that intensified FTC did not lead to significant losses of winter N resources (e.g., abstract line 36; discussion lines 435, 452, 460–465) is therefore not fully supported by the data, as the study lacks direct measurements of N loss pathways such as leaching or gaseous emissions (e.g., NO₃⁻ leaching, denitrification, or volatilization). It would be valuable for the authors to clarify the scope and limitations of the ¹⁵N tracer method in assessing winter N retention, explicitly acknowledging that it only tracks added N and does not capture mobilization and potential loss of native soil N.

4. Plant-microbe interactions:

The manuscript interprets the temporal pattern of stable microbial ¹⁵N retention and increasing plant ¹⁵N uptake as evidence of a decoupling mechanism that stabilizes winter N resources and fosters mutually beneficial plant-microbe interactions under intensified FTC. However, there is no observed decline in microbial ¹⁵N over time (Figure 6), which would be expected if immobilized N were later released to support plant uptake. Furthermore, the dynamic appears consistent across all treatments, suggesting it is a general feature of seasonal nitrogen cycling rather than a specific effect of FTC. Therefore, the conclusion that FTC induces a stabilizing mechanism through temporal decoupling is not directly supported by the data and should be removed.

5. Clarify scope of conclusion regarding FTC effects:

The discussion (line 454) mentions that intensified FTC increased total ¹⁵N recovery, and the conclusion (line 538) states that intensified FTC reveals enhanced retention of winter N resources. However, the data show that this effect is limited to the high-frequency FTC treatment, with no comparable increase under low-frequency FTC. Therefore, the current phrasing could be misinterpreted as evidence of a general ecosystem response or resilience to FTC. Please clarify that the observed effect pertains specifically to high-frequency FTC.

6. Temporal resolution of sampling and N cycling processes:

While the seasonal soil sampling intervals are appropriate for tracking broad patterns, the temporal resolution immediately before, during, and after the freeze-thaw treatments appears too coarse to capture short-term N cycling processes. Processes like nitrification and denitrification usually occur within days after FTCs and can contribute to substantial N losses in form of N₂O fluxes. Please consider acknowledging this limitation, especially when interpreting mechanistic effects of FTC on nitrogen transformations.

7. Soil sampling depth and deep roots:

The study sampled only the top 20 cm of soil, which likely captures much of the microbial and plant root activity. However, nitrate is highly mobile and may leach below 20 cm, especially following FTC-induced mineralization, potentially leading to underestimation of N losses and overestimation of retention. Moreover, Hypothesis 2 suggests that deep-rooted species may increase winter N uptake under intensified FTC. Yet, root ¹⁵N retention was assessed only within the top 20 cm, which may not reflect uptake from deeper soil layers where such species may access nitrate. This could reduce the apparent contrast between shallow- and deep-rooted species. We recommend acknowledging these limitations when interpreting both N retention dynamics and species-level uptake patterns.

8. Species specifics (Hypothesis 2):

You hypothesize that “intensified FTC would lead to differential utilization of winter N sources among plant species,” and in the discussion, you refer to several plant traits to explain species-specific responses. However, it is currently difficult for the reader to follow these arguments without a clear overview of the relevant species and their functional traits. To strengthen the link between your hypothesis and interpretation, we recommend including a brief summary of the observed species and their key traits that you mentioned in your hypothesis 2 (competitive abilities, root system architecture (particularly rooting depth and winter root activity), temporal niche partitioning in growth phenology (early spring green-up), and susceptibility for root damage) in the Methods section. Further, we suggest to formulate the hypothesis more concretely. This would provide helpful context for understanding the mechanisms underlying your findings.

9. Site specifics:

The manuscript provides useful site-specific information; however, the methods section does not specify how these data were obtained—whether from field measurements, previous studies, databases, or modeling (e.g. bulk density). Including details on the sources and collection methods for these site parameters is important for transparency and reproducibility. Moreover, while site differences are documented, the manuscript lacks a discussion of how these environmental and edaphic differences may have influenced the observed results. Since the study compares two distinct sites, integrating an analysis or interpretation of how site characteristics might drive differences in nitrogen cycling, microbial activity, plant uptake, or freeze-thaw responses would strengthen the ecological context.

10. Meteorological information:

The manuscript references the China Meteorological Administration website (http://data.cma.cn/) as the source for meteorological information at the study sites. However, this website is primarily in Chinese and can be difficult to navigate for non-Chinese speaking readers. To improve accessibility and reproducibility, we recommend providing a more direct and specific link to the exact data pages used, or alternatively, suggesting an English version or database where the meteorological data can be accessed more easily by an international audience. This would help readers verify the data and facilitate broader use of the study’s findings.

11. Artificial manipulation and microclimate measurements:

The study aims to address the lack of natural in situ FTC experiments (lines 71), but the use of polyester tents and air heating still introduces some artificial influences. While the manuscript notes that mesh windows were used to reduce CO₂ accumulation, it would still be helpful to clarify how other potential microclimatic changes were accounted for. For instance, air temperature was recorded at 5 cm above ground (line 174), but these data are not presented or discussed. Changes in humidity, photosynthetically active radiation (PAR), or CO₂ levels could still affect plant growth and nitrogen uptake. We recommend briefly discussing these possible side effects of the experimental setup to help contextualize the findings. Also, figure 2 shows that the soil processing period overlaps with the occurrence of freeze-thaw cycles at the sandy steppe site. Please address if this overlap may have influenced the results.

12. Soil moisture measurement and implications of differences:

While soil moisture data were recorded using data loggers (line 205), the manuscript does not specify the sensor types, calibration methods, or how values (in m³ m⁻³) were derived. The presence of negative soil moisture values suggests possible measurement or calculation errors that should be addressed. Additionally, there is an inconsistency between the text (lines 280–284), which states that elevated soil moisture occurred only in early spring, and Figure 2b, which appears to show sustained increases under LFTC and HFTC throughout much of the season. It should be included in the discussion how treatment-induced changes in soil moisture may have influenced nitrogen dynamics and plant ¹⁵N uptake, as it was a significant predictor in the correlation and random forest analyses.

13. Restructure Chapter 2.3 Sampling and Processing:

This section would benefit from restructuring to more clearly distinguish which subsamples were used for which analyses and to detail the analytical procedures more consistently. We would recommend to reorganizing this section to clearly present the workflows for each measured variable (e.g., soil mineral N, DOC, microbial C and N, soil/plant/microbial ¹⁵N, soil moisture and temperature) to improve readability and reproducibility. Further, it is unclear whether the described K₂SO₄ extraction and analysis refer only to microbial biomass C and N samples or if the same procedure was used for soil mineral N (NH₄⁺ and NO₃⁻) and DOC analyses as well (lines 231-233). Please clarify whether different extraction procedures were used for mineral N and DOC, and if so, provide the details (e.g., solution type, shaking duration, soil-to-solution ratio). Regarding the ¹⁵N measurements, plant and soil ¹⁵N were measured with an elemental analyzer coupled to IRMS — was the same system used for total microbial ¹⁵N, or was a different method used? Please also clarify whether the same elemental analyzer (Elementar Vario Max CN) was used for all C/N and ¹⁵N analyses, or if multiple instruments were involved.

14. Correlation analysis and random forest:

The manuscript includes correlation and random forest (RF) analyses to identify key predictors of plant ¹⁵N acquisition, incorporating variables such as DOC, microbial carbon, and microbial community composition (bacterial vs. fungal biomass). However, the rationale for including both statistical approaches is not clearly explained, and the results from these analyses are not sufficiently integrated into the discussion. It remains unclear how the outputs of both approaches complement each other, and what ecological insights they offer regarding nitrogen dynamics under freeze-thaw conditions. Additionally, the manuscript does not explain the relevance of DOC and microbial biomass C to nitrogen cycling or FTC effects. Similarly, the ecological importance of differentiating microbial groups (bacteria vs. fungi) is not introduced and not interpreted in the results or discussion. Lastly, it is unclear why these analyses were performed only for the two FTC treatments and not for the control. To improve coherence, we recommend explaining the rationale for including both correlation and RF approaches and discussing the ecological relevance of the identified predictors.

15. Discussion on the relevance and magnitude of observed differences:

While the manuscript highlights statistically significant differences between treatments, it does not sufficiently address the ecological or functional relevance of these differences. A more in-depth discussion is needed on the magnitude of the changes observed. Please consider discussing whether the observed changes are likely to have substantial impacts on grassland resilience, nutrient cycling, or plant community structure in the context of winter climate change.

16. Future research directions and limitations of the study:

The future research section outlines useful directions, particularly regarding microbial functional traits and long-term ¹⁵N fate. However, other potentially impactful avenues are overlooked. For instance, it would be valuable to differentiate between the origins of newly available nitrogen during freeze-thaw cycles — specifically, whether it stems from microbial cell lysis, root mortality, or physical disruption of soil aggregates. Distinguishing these sources could significantly enhance mechanistic understanding of nitrogen retention and loss pathways. Additionally, the study briefly references nitrogen losses but does not address gaseous emissions. Monitoring greenhouse gases (e.g., N₂O, CO₂) during FTC events could offer further insight, especially given that FTC induced N₂O peaks often occur without corresponding CO₂ increases, which could have a relation to your results e.g. on increases in microbial biomass N and decreases in microbial biomass C.  

Technical Corrections

Line 273-395 Results: We suggest removing redundant “p < 0.05” and instead define significance threshold in the Methods or where relevant, report actual p-values.

Figure 2: We recommend to make the figure broader to better see the single FTCs. The labeling of the dates is unclear. The legend for samplings looks like another sampling event which is confusing.

Figure 3, 4 and 6: The dots for single measurements are not necessary and just confusing if they overlap – having error bar should be enough.

Figure 5 and 7: Both are exactly the same plot, figure 5 doesn´t fit to the description – the figure showing plant biomass N is missing.

Figure 3 to 7: please including the exact dates of sampling in the legend.

Introduction line 37: replace “while” with “While”

Line 106: does -2-1°C mean -2.1 °C?

Line 125: replace “the predominant soil type in meadow grassland is loam soil, and which in sandy grassland is sandy loam soil.” With “the predominant soil type in meadow grassland is loam, while in sandy grassland it is sandy loam.”

Line 131: add space “78 %”

Line 161: There is a verb missing: “, and no significant differences in plant/microbial N concentrations when compared to the 15N treatments.”

Line 172: add space “15 cm“

Line 231: add space “60 ml“

Line 232: How much Molar was the K₂SO₄ solution?

Line 232: add space “30 min”

Line 234: please state for what the conversion coefficient is used.

Line 244: replace “as well as microbial community structure in situ soils” with “as well as the microbial community structure of in situ soils”

Line 250: The part with the calculation of ¹⁵N acquisition/recovery is not statistical analysis.

Line 253: replace “;” with “,”

Line 258: add space “x V x”

Line 259: start new paragraph

Line 264: Is “rcorr” from R or SPSS?

Line 266: randomForest and rfPermute packages from R or SPSS?

Line 271: remove comma “SigmaPlot 14.0”

Line 271: Origin 14.0 is not existing

Line 335: replace “⁵N” with “¹⁵N”

Line 377: add space “C and” and replace “Under” with “under”

Line 388: replace “plants” with “plant”

Line 400: subsequent growing seasons would mean several years as there is one growing season per year

Line 414: Abbreviation of MBN was already introduced

Line 421: remove “.”

Line 436: specify that you mean plant water uptake

Line 457: A verb is missing: “indicating that effective ecosystem-level N retention mechanisms.”