Reviewer comments #1 (RC1) - Chantal Hendriks:

With pleasure I read the manuscript entitled ‘Managing Soil Nitrogen Surplus: The Role of Winter Cover Crops in N2O Emissions and Carbon Sequestration’. There are still quite some knowledge gaps to fill regarding this topic and the authors contributed to bridging some of these gaps. In general, I think the manuscript is publishable after some minor revisions. Especially the carbon modelling needs more explanation. More specific comments are listed below.

RC1: L26-30: the field trial was only 16 months, but you estimated carbon sequestration over a 50 year time period. I guess the authors used simulation models to assess the potential long-term sequestration. But this should be added to the abstract.

AR: Thank you for your comment. To clarify, we have now added two sentences to the abstract (L19-21) to explicitly state that estimates of soil C sequestration were derived by coupling ensembles of soil organic C models with observed C inputs from catch crop biomass. This addition ensures that readers are informed up front about the methodological approach used for long-term sequestration projections.

RC1: L31-33: the authors recommend “optimized cover crop selection”, but according to the results, not much difference in N2O emission and C sequestration is noticed between the different cover crop varieties.

AR: Thank you for pointing this out. The primary aim of our statement in L31-33 is to emphasize the need for further research to quantify and optimize the effects of catch crops on land-use-related greenhouse gas emissions in dynamic agroecosystems. While our results show relatively low variability in soil carbon sequestration among the tested catch crops at the two sites, this may be due to the relatively low carbon inputs from the catch crop biomasses, likely caused by late planting dates, as discussed in Section 4.5.2. However, despite this limited variability, we observed some differences. For instance, C sequestration rates associated with spring vetch were approximately 25-40% lower compared to winter rye (Table 5). This highlights the potential for mitigating greenhouse gas emissions through optimized catch crop management.

Regarding N₂O emissions, our findings show that frost-sensitive cover crops emitted more N₂O during the growing season in trials with harsher winter frosts compared to trials with milder winters. This highlights the importance of considering local climatic factors, such as the likelihood of frost events, when selecting cover crop varieties. These results suggest that while differences in the effects of cover crops may appear subtle, careful selection and management of cover crops, tailored to site-specific conditions, could enhance their contributions to greenhouse gas mitigation.

RC1: L75-77: this is a stand-alone statement which comes out of the blue. Elaborate on it (because the authors also focused on the short-term N2O emissions), or delete the sentence (or move it to the discussion).

AC: Thank you for this suggestion. We have moved the sentence to the discussion section, where it now serves as an opening to a paragraph emphasizing the importance of long-term N₂O measurements. This provides better context for understanding the full impact of cover crops on N₂O dynamics and aligns more closely with the focus of the discussion.

RC1: L81: “soil organic models”, I think the word ‘carbon turnover’ or ‘matter turnover’ is missing here.

AR: Thank you for pointing this out. We have replaced the term with “carbon turnover model” in the revised manuscript to clarify the intended meaning.

RC1: L109: I’d recommend to describe the soil characteristics for each Luvisol separately, so a reader knows which soil has a soil organic matter content of 20 g/kg and 30g/kg.

AC: Thank you for the suggestion. We have added a new table (Table 1) to the revised manuscript, which specifies the soil properties and site information for each trial. This provides a clear and detailed comparison of soil characteristics across the study sites. The text has been revised accordingly to prevent repetition.

RC1: L115-135: it would help if the experimental design is accompanied with a table.

AC: Thank you for the suggestion. The management practices of the various crops, along with details of the experimental design, are now outlined in Table 2 to provide a clearer and more comprehensive overview for readers. The text has been revised accordingly to prevent repetition.

RC1: L154: please add a reference to your assumption of 2.65 g/cm3.

AC: Thank you for pointing this out. A reference has been added in the revised manuscript to support the assumption of 2.65 g/cm³.

RC1: L156: please explain why you used two types of chambers. Can the results still be compared, because the volume of the two chambers differs?

AC: The use of two different chamber types was necessary for practical reasons. The large round chambers (60 cm diameter) used for cover crops and winter wheat were not suitable for sugar beet plants, as the plants would have outgrown the chambers in the later stages of cultivation. Vice versa, the rectangular chambers wouldn’t have worked for the cover crops and winter wheat as the row spacing is not large enough. Therefore, rectangular chambers were used for the sugar beet phase, positioned between the rows of plants to accommodate their growth and round ones that include the plants in their circumference for the other phases. The results from the two types of chambers are comparable because the calculation of N₂O emission rates takes into account the chamber measurements (height and surface area). In addition, the calculation method we used (package gasfluxes in R) includes adjustments for deviations from linearity in the flux rate, handling cases where the emission rate forms a saturation curve (Fuss and Hueppi, 2020). This ensures the reliability and comparability of the results despite the differences in chamber design.

RC1: L159: add ‘N2O fluxes’ between ‘measured and ‘using dark’. Again, how can you compare these results with the results from the other chambers used in Gottingen?

AC: Thank you for your comment. For clarification, the term ‘gas’ has been added to the text before ‘fluxes’ to reflect that we measured both CO₂ and N₂O. In addition, the word ‘dark’ has been removed as this applies to both sites and is standard for GHG measurements. Regarding the comparison of results between chambers, the rationale given in response to the previous comment applies here as well. The calculation of N₂O flux rates takes into account the differences in chamber volume as well as non-linearity in increased gas concentrations, thus ensuring comparability of results between the different chamber types.

RC1: L172: replace (IPCC, 2019) for IPCC (2019)

AC: Thanks for pointing this out. The citation has been changed to IPCC (2019) as suggested.

RC1: L201: How do you know the effect in C stock change is caused by the addition of cover crops when you also apply other organic inputs (30m3 digestate) at maize?

AR: Thank you for this insightful question. We quantified the effect of cover crops by comparing a scenario with cover crops (either saia oat, winter rye, or spring vetch) to a control scenario without cover crops. Importantly, the management practices for the control and cover crop scenarios were otherwise identical, including the application of digestate in crop rotation CR1. This approach minimizes the influence of digestate application on the observed differences in C stock changes between scenarios. However, it is important to note that cover crop cultivation can influence decomposition rates, as seen in some models of the ensemble (e.g., RothC). This interaction between cover crop effects and other management practices (such as digestate application) reflects dynamics that are likely to occur under real-world conditions. The methodology for quantifying these effects is described in Lines 205-208 (preprint). To enhance clarity, we have added a sentence to explicitly state that the modeling included a control scenario without any cover crop (L201-203 in the revised manuscript).

RC1: L206: I miss some information on the modelling using RothC and C-Tool. At the moment I’m not able to reproduce your modelling exercise. How did you use both tools (e.g., in an ensemble run or did one model complement the other)?

AR: As described in L209-220, RothC and C-Tool were applied as part of a model ensemble with three ensemble members, following the approach outlined in Seitz et al. (2023). The results presented in the manuscript represent the average and variability across the ensemble members, providing robust estimates of soil carbon sequestration.

RC1: How did you initialize the SOC stock?

AR:  For RothC, we assumed equilibrium conditions for initializing SOC stocks. The distribution of SOC pools corresponded to this equilibrium assumption, and we used an analytical solution of the RothC model to quantify the initial pool distributions based on the initial SOC stocks, as described in Seitz et al. (2023). For C-Tool, we followed the initial model fractions described by Taghizadeh-Toosi et al. (2014), which detail the initialization process for simulating whole-profile carbon storage in temperate agricultural soils. To enhance clarity, we have added a sentence and edited the paragraph in the revised manuscript to better describe the model initialization process.

“Model initialization was performed separately for each model in the ensemble. For RothC, pool distribution at equilibrium was determined using an analytical solution (Dechow et al., 2019), while for C-Tool, the initial pool fractions followed the approach proposed by Taghizadeh-Toosi et al. (2014).”

RC1: What historical management took place on the fields?

AR: The experimental plots were managed as cropland prior to the study. Historical management practices included the cultivation of typical crop rotations such as winter wheat, winter barley, sugar beet, and oilseed rape. Soil management practices consisted primarily of ploughing and conventional tillage. Organic fertilizers such as digestate were used in some years, particularly for sugar beet and corn crops, at varying rates. In addition, lime products were used to amend the soil in some years. This historical management provided a baseline of standard farming practices.

RC1: Which site-specific input data did you require/use and which input data did you assume (e.g., soil depth, climate data, cover factor (what assumption did you make)?

AC: The following site-specific input data and assumptions were used for the modeling:

    Soil Depth: SOC sequestration was simulated for the 0-30 cm soil horizon, as now specified in L202 of the revised manuscript.

    Climate Data: Monthly weather data (precipitation, temperature, and global radiation) for the period 2018-2021 was obtained from DWD grid data (DWD, 2023). To extend the data to cover the 50-year simulation period, these time series were repeated.

    Crop Cover Factor: The crop cover factor, used in the RothC model, was set according to the guidelines provided by Coleman and Jenkinson (1996). Specifically, a factor of 0.6 was applied to months where crop cover was predominant, while a factor of 1 was used for months dominated by bare fallow.

RC1: Is there any irrigation in the fields, or ploughing?

AC: No irrigation was applied to the plots during the experiments. Ploughing was performed within the simulation depth of 30 cm; however, it did not affect the modelled carbon stocks from the perspective of the RothC and C-Tool models. This is because both modeling approaches do not consider the effects of ploughing on the dynamics of SOC decomposition.

RC1: What are the soil properties?  

AR: To avoid redundancy, the soil properties are described in detail in Section 2.1, "Study Sites and Experimental Design." Additionally, we have now added a table (Table 1) that provides a clear overview of the soil characteristics for the different trials.

RC1: Why choose for this model and not for a model that assesses C and N fluxes?

AR: We chose ensembles of soil organic carbon models because they have been extensively validated for describing management-induced soil organic carbon changes in long-term field experiments and observational sites in Germany (e.g., Dechow et al., 2019; Riggers et al., 2019). These models have proven reliable for assessing SOC dynamics under various management practices.

To our knowledge, similar extensive model evaluations do not exist for more sophisticated C-N models for German croplands. Biogeochemical models, while capable of assessing both C and N fluxes, require significantly more detailed input data at higher temporal resolutions, which is often unavailable for long-term experiments. There are model extensions of SOC models, such as the "sorcering" package (Scherstjanoi and Dechow, 2021), that enable quantification of N mineralization. However, in these approaches, N cycling does not influence C cycling, meaning the expected results for C cycling would align closely with those of the SOC models used in our study.

RC1: L217: the source you refer to studied tree species. How applicable is this approach for green manure and more specifically to the green manure types that were included in this study?

AR: While the referenced formula was originally developed for tree species (Gale and Grigal, 1987). it has since been applied to a variety of ecosystems, including croplands and grasslands, as demonstrated by Jackson et al. (1996). In their study, Jackson et al. provided ecosystem-specific parameters for croplands, which we adopted to model root distribution with depth in our study. This clarification has been included in the text, along with an additional reference (Jackson et al., 1996).

RC1: L219: replace ‘a parameter’ for ‘a plant-specific parameter’

AR: Thank you for the suggestion. Since we describe the calibration of parameters specific to the observed catch crops (as per Equation 3), we agree that "plant-specific" is appropriate. We have updated the text to use the term "plant-specific values of β" instead.

RC1: L226: why did you decide to copy the weather data 2018-2021. These were extremely dry years and might not be representative on the long term (as also mentioned in L230). Consider climate scenario’s or a longer time range.

AR: It is correct that the years 2018 and 2019 were extremely dry years, but these were the conditions under which the field experiments were conducted, and the observed cover crop biomass (saia oat, spring vetch, and winter rye) was recorded. Since these biomass values were used to estimate carbon inputs for our scenarios, we chose to maintain consistency between the observed cover crop biomass and the corresponding weather conditions. As a result, our approach does not aim to capture representative long-term climate conditions but rather to extrapolate the effects of cover crops on soil organic carbon sequestration under the meteorological conditions observed during the field experiments. This is also consistent with the N₂O fluxes and SMN levels reported in this manuscript, as these results are based entirely on field experiments conducted between 2018-2021.

Incorporating climate scenarios or longer time series would require dynamic modeling of cover crop development in response to meteorological conditions. However, to our knowledge, there is currently no biogeochemical model calibrated to simulate the growth of the specific cover crops studied (saia oats, winter rye, and spring vetch) and their effects on SOC stock changes under German weather conditions. The implications of using experimental field data to model the effects of cover crops on SOC sequestration are discussed in detail in Section 4.5.2.

RC1: L250: it is not clear to me how the site-specific weather data differ from the DWD weather data. Also explain in Chapter 2.5 why you used DWD instead of site-specific weather data. I agree with your decision, but it might cause some confusion.

AR: Thank you for your comment. We used DWD weather data as model input to ensure a consistent and complete climate dataset for the relevant simulation period (January 2018 to December 2021). This choice was made because the DWD dataset spans the entire experimental crop growth period at both sites, providing a consistent basis for comparison between sites. Furthermore, the suitability of DWD monthly grid data for modeling applications has been demonstrated in studies evaluating soil organic carbon models at German permanent observation sites (e.g., Riggers et al., 2019). For clarification, we have added this explanation in Chapter 2.5 to avoid potential confusion about why site-specific weather data were not used.

RC1: L285: the author did not mention the N fertilization of sugar beets before. This should be added to the methodology.

AR: Thank you for pointing this out. The N fertilization for sugar beets is already mentioned in the preprint in Lines 130-133 (Section 2.2). To enhance clarity and accessibility, we have now included a detailed table of field management practices (Table 2 in the revised manuscript), which explicitly outlines the N fertilization rates and dates, as well as management dates for each trial. This addition ensures that all relevant information is centralized and easily accessible to readers.

RC1: L321: why did vetch show N2O peaks, and why did only G18 show peaks and G19 not?

AR: The N₂O peaks observed for vetch in G18 can be attributed to the severity of frost events and the resulting damage to the cover crops, as described in Section 3.2 (Lines 253-257 in the preprint). In G18, vetch was completely terminated by frost damage, leading to the release of N and subsequent N₂O emissions. In contrast, G19 experienced a milder winter, resulting in only partial frost damage to the vetch, which likely reduced the extent of nitrogen release and related N₂O emissions.

RC1: L415: the text below and the figures do not match. I’d expect two scenario’s, one for CR1 and one for CR2, and the baselines (controls). Please, clarify the modelling approach.

AR: As mentioned before effects of cover crops on soil carbon stocks were represented by the difference between cover crop scenarios and the control without scenarios and this is shown in figure 4. We modified the captions to be more clear: “Modelled effect of cover crops on the increase in SOC stocks (0-30 cm) compared to the control without cover crop for regionally common crop rotations…”

RC1: L430: linking the results to research done in a completely different climatic zone requires more explanation or needs to be removed.

Thank you for pointing this out. The linkage to semi-arid regions was based on the observed annual rainfall of 430 mm in Göttingen in 2018, which is within the range of precipitation typically associated with semi-arid areas (250-500 mm annually). To clarify, we have revised the text to explicitly highlight that the rainfall conditions in 2018 were uncharacteristically low for the region and comparable to semi-arid environments, where annual rainfall typically ranges from 250 to 500 mm.

RC1: L445: in Chapter 4.2 some results are mentioned. Consider combining the Results and Discussion section or move the results to Chapter 3.

AR: Thank you for pointing this out. In the revised manuscript, we have deleted the repeated results to avoid redundancy. We are also considering combining the Results and Discussion sections to enhance the flow of the manuscript. However, this decision will depend on whether it aligns with the journal's guidelines.

RC1: L622-624: do not repeat the results

AR: thank you for pointing this out, we have deleted repeated results in the revised manuscript.

RC1: L635 - 644: do not repeat the results. Re-write this section and try to be more concise.

AR: thank you for this suggestion; we have shortened the discussion of C modelling in the revised manuscript.

RC1: Due to the high number of hypotheses, the Discussion is exhaustive and good, but extremely long. Perhaps consider a restructuring and start with an overview of the hypotheses (rejected or accepted) followed by a discussion and underpinning of the results for each hypothesis

AR: Thank you for this valuable suggestion. We will carefully consider ways to improve the structure and conciseness of the discussion in the revised manuscript.

Reviewer comments #2 (RC2): The present manuscript undertakes the estimation of the impact of multiple cover crop species on direct and indirect N2O emissions, and C sequestration. The paper addresses these aspects in the context of a crop rotation that goes beyond the cover cropping phase (including sugar beet and winter wheat), a relevant aspect when studying the impact of this practice. It also considers multiple variables for a more comprehensive assessment of environmental and productivity outcomes of cover cropping. I consider this paper is suitable for publication with minor revisions, which I share below. Congratulations on this great work!

RC2: Title: would it be possible to more clearly reflect the geographical and methodological scope of the work conducted in the title of the paper? The current title may lead readers to think this paper is a comprehensive review on the effect of cover crops on N2O emissions and SOC sequestration across multiple locations, growing conditions, practices, etc.

AR: Thank you for this valuable suggestion. We understand the concern that the current title might suggest a broader review rather than the field-based study we conducted. To address this, we propose the following alternative title:

"Evaluating N₂O Emissions and Carbon Sequestration in Temperate Croplands with Cover Crops: Insights from Field Trials."

This title more clearly reflects the geographical focus (temperate croplands), the methodological scope (field trials), and the main outcomes assessed (N₂O emissions and carbon sequestration). We believe it provides more clarity to potential readers and aligns with the content and focus of our manuscript.

RC2: Line 19: consider adding a few sentences in the abstract that describe how the experiment was set up.

AR: Thank you for pointing this out. We have revised the abstract to include a brief description of the experimental setup (lines 19-21 in revised manuscript).

RC2: Lines 75-77: could this sentence be included in another paragraph in order to avoid having a paragraph with a single sentence?

AC: Thank you for highlighting this issue. As mentioned above, we have addressed it by moving the sentence from lines 75-77 to the discussion section in the revised manuscript. It now serves as the opening to a paragraph emphasizing the importance of long-term N₂O measurements. This relocation enhances the flow of the manuscript and provides better context for understanding the impact of cover crops on N₂O dynamics. The revised placement also aligns more closely with the discussion's focus on the implications of our findings.

RC2: Line 81: what do you mean by “soil organic model”? Should that be “soil organic C model”?

AR: Thank you, we replaced the term by “carbon turnover model” in the revised manuscript.

RC2: Lines 103-104: would it be possible to add the geographic coordinates of the sites?

AR: Thank you for that suggestion. We have included the geographic coordinates of the experimental sites in a new table (Table 1) in Section 2.1 of the revised manuscript. This table provides site-specific details to ensure that readers can easily reference the geographic context of the study.

RC2: Line 131: Is the total of 180 kg N/ha the addition of soil N plus fertilizer N? Would it be possible to clarify that? Were all site*year combinations sampled the same way in the spring for SMN? Also, could the German fertilizer ordinance be cited?

AR: Thank you for your comment. The total amount of 180 kg N ha^-1 represents the sum of SMN levels measured in the 0-90 horizon in March and the additional mineral N applied as fertilizer. SMN sampling in March was conducted consistently across all site-year combinations, as described in Section 2.2. The text has been revised for clarity, and a citation to the German Fertilizer Ordinance (German Fertilizer Ordinance, 2017) has been added. Fertilization rates and dates are detailed in Table 2 of the revise manuscript to further clarify the methodology.

RC2: Line 146: Why were the first 3 cm of soil not considered in the measurements of bulk density? Also, why were the soil layers considered for bulk density measurements not continuous? There seems to be a gap between the 1st and 2nd layer, as well as between the 2nd and 4th layer.

AR: We selected the sampling depths as representative for the 0-–10, 10-20, and 20-30 cm layers, covering the topsoil. Continuous bulk density measurements with 5 cm increments (e.g., 0-5, 5-10, 10-15 cm, etc.) would theoretically provide greater accuracy. However, this approach is not feasible with the applied method due to the disturbance caused during sampling, which affects the surrounding soil. To maintain accuracy and consistency, we introduced buffer zones and sampled the central 5 cm of each 10 cm layer, assuming this adequately represents the entire layer. The first 3 cm of the soil were not considered because surface soil is influenced by numerous external factors (e.g., organic residue, surface sealing) that are not directly representative of our treatments. This methodological choice ensures more robust and reliable results for the studied layers.

RC2: Line 149: should g/cm3 be g cm-3? Line 152 seems to have similar opportunities to refine the notation used.

AR: Thank you for pointing this out. We have updated the notation throughout the manuscript to use "g cm^-3" consistently.

RC2: Section 2.2: were soil samples for SMN and WFPS taken every time that N2O was measured?

AR: Gas samples for N₂O measurements were collected weekly (or more frequently during specific events), while soil samples for SMN and WFPS analysis were collected biweekly, as described in Section 2.2. This frequency was chosen because soil dynamics generally change less rapidly than N₂O emissions.

RC2: Section 2.3: how many chambers per plot were there in the study? One? Consider clarifying. Also, were gas samples taken at specific time intervals? Consider clarifying. Also, consider indicating how many times N2O was sampled across the site and years. Consider including similar information on the number of samples taken for other variables, like SMN, WFPS, etc.

AR: There was one chamber per plot for gas sampling. Gas samples for N₂O measurements were collected at 20-minute intervals. These details have been clarified in Section 2.3. Gas samples for N₂O measurements were collected weekly, while soil samples for SMN and WFPS were collected biweekly for the entire 16-month trial period. We have opted not to include the exact number of samples, as it varies slightly between sites and years, and does not significantly affect the interpretation of results.

RC2: Line 177/178: should “...normal distribution of data…” be “normal distribution of model residuals…” instead?

AR: Yes, thank you for pointing that out. We have corrected this in the revised manuscript.

RC2: Line 191: extra comma after 2015.

AR: Thank you for noticing this. The extra comma after "2015" has been removed in the revised manuscript.

RC2: Line 197-198: is the application of digestate a common practice in the region, or why was it included?

AR: Yes, the application of digestate, particularly in combination with silage maize cultivation, is a common practice in the region.

RC2: Line 197: what does “cover crop/bare fallow” in CR1 and CR2 mean exactly? Could that be clarified?

AR: We have clarified this in the revised manuscript. CR1 and CR2 refer to control treatments with bare fallow, while the cover crop scenarios involve the inclusion of cover crops such as saia oat, spring vetch, and winter rye (Lines 208-210 in the revised manuscript).

RC2: Line 201: consider clarifying that by “treatment” you mean the cover crop treatments tested in the field experiments. In the same line, consider replacing “have been quantified” for “were quantified”.

AR: Thank you for this suggestion. We have replaced “have been quantified” with “were quantified” as suggested. Additionally, the sentence mentioning “treatment” has been removed for clarity.

RC2: Line 203: is this first order kinetic modeling approach the one used by RothC and C-Tool? Please consider clarifying, or better integrating the paragraphs to more clearly explain how the modeling was conducted (207:208)

AR: The first-order kinetic modeling approach used in this study differs from RothC and C-Tool. We have clarified this distinction and better integrated the explanation of the modeling methodology in the revised manuscript.

RC2: Line 212: should “variant-specific” be “variety-specific”?

AR: We have revised the sentence to provide a more detailed explanation. “Yields of main crops after the cover crop (cover crop scenarios) or after the alternative bare fallow period (control) were based on observed yields from experimental treatments where the main crops followed the same cover crops (saia oat, spring vetch, or winter rye) or bare fallow treatment”.

RC2: Lines 225-227: does this mean that four years of weather data (2018-2021) were repeated for 50 years to conduct the simulation? If that was the case, how representative the weather during those 4 years is from general climate patterns in the areas of the study? I think a larger time period should probably be considered for a 50-year simulation. I think this point should definitely be reevaluated. Furthermore, should future climate scenarios consider the impacts of climate change (change in temperature and rainfall patterns, for example)?

AR: As mentioned above, we chose to repeat the weather data from 2018 to 2021 because the carbon inputs from cover crops used in the simulation were based on observed above- and belowground biomass for cover crops measured at the study site, under the specific meteorological conditions during these years. Similarly, carbon inputs from the main crops following the cover crops (sugar beet and winter wheat) were quantified using observed aboveground biomass and yield data from the field experiments. This approach aims to quantify the potential effects of cover crops on soil carbon sequestration under the specific meteorological conditions of the field experiments. The same conditions were used for N₂O flux and SMN levels, ensuring consistency with the experimental data.

We acknowledge the concern about using a larger time period for a 50-year simulation. However, the focus of this simulation is on the observed conditions specific to the study period. Regarding climate change impacts, future simulations could incorporate changing temperature and rainfall patterns; this will be considered in further research to provide a broader context for long-term predictions.

RC2: Lines 234 and 235: seems like soil temperature was measured, but I do not think that is listed in the materials and methods section. Could that be added?

AR: Thank you for your observation. Soil temperature measurements were indeed taken during the study. In the revised manuscript, we have updated the sentence for clarity: “Meteorological data, including daily precipitation and hourly soil and air temperatures, were collected from stations located at the field sites.” This ensures that soil temperature measurements are explicitly mentioned in the Materials and Methods section.

RC2: Line 250: how were the crop phases defined? Do they go from planting until harvest? Consider clarifying somewhere in the methods.

AR: Thank you for your comment. We have added a clarification at the end of Section 2.1: “Crop phases for the cover crops and sugar beet were defined from sowing to harvest, except for winter wheat, for which the crop phase ended at the first fertilization date.”

RC2: Lines 261, 262: I imagine N uptake was estimated based on DM yield and N content, correct? If that is the case, should 100% of the N content estimated for vetch be considered N uptake? Or is part of that N coming from N fixation? Please revise and clarify.

AC: Thank you for this important comment. We have revised Section 3.2 to clarify this: “N uptake was estimated based on DM yield and N content. For vetch, the N content includes both N taken up from the soil and N derived from biological nitrogen fixation.”

RC2: General methods: was sugar beet and winter wheat yield measured? In case yes, was there any impact of the treatments? If yes/no, consider including a few sentences in the results and discussion sections related to these points.

AR: Yes, sugar beet and winter wheat yields were measured. The results concerning sugar beet yields have been published by Koch et al. (2024). We opted not to include these results in the current manuscript, as they extend beyond the primary scope of this study, which focuses on N₂O emissions, soil C sequestration, and SMN dynamics.

RC2: Lines 262, 263: you indicate “A strong correlation was observed between cover crop DM formation and both C content and N uptake (r>0.9).” Is this because of the way in which N uptake is calculated?

AR: Yes, to improve accuracy, we have revised the sentence to replace “C content and N uptake” with “C and N contents.”

RC2: Lines 291, 292: you indicate: “At the end of the cover crop phase in April, SMN levels were lowest in rye, with some variation in other treatments across site-years (Figure 1, Table S2).”; are the values in figure 2 averages for the entire phase, or values associated with sampling at the end of each phase? The figure caption suggests it is the average across the entire phase. Please clarify.

AR: The values in Figure 2 represent the average SMN values for each phase, as indicated in the figure caption. The text in Lines 291-292 refers to Figure 1, which shows the dynamic SMN values for each sampling date. Table S2 also contains the SMN values for specific time points.

RC2: Section 3.5: I really liked this section! In my opinion, it is super useful for the reader to have this recap!

AR: Thank you for your positive feedback! We are glad you found this section useful and informative.

RC2: Figure 2 and 3: would it help improve the clarity and comparability across panels to use the same Y axis scale for all panels, or at least for panels a, b and c?

AR: Thank you for this valuable suggestion. While we initially chose different Y-axis scales to better emphasize the differences across treatments, we will revisit this choice in the revised manuscript to assess whether using a consistent Y-axis scale (or a uniform scale for panels a, b, and c) enhances clarity and comparability

RC2: Table 3: why was each cover crop specie included as a separate fixed effect in the linear mixed-effect model? Should cover crop be included as a categorical fixed effect, with three levels (one for each of the three species)? What is the impact in your model of including them separately? Also, seems like different models were fitted for each crop phase, correct? Consider clarifying this in the methodology.

AR: Thank you for your detailed question. We did include cover crop treatments as a categorical fixed effect with three levels (one for each species). The same model structure was used for all phases, but datasets were subsetted for each phase. Section 2.5 has been revised to clarify these points.

RC2: Table 3: conditional R2 describes model fit considering both fixed and random effects. Please consider calculating marginal R2 that consider only model fixed effects.

AR: Thank you, we will consider this for the revised manuscript.

RC2: Table 3: how was the number of observations for each phase calculated? Also, were soil samples for SMN taken every time that N2O was measured? It is not clear from the methodology, and this modeling approach assumes there is SMN, WFPS and soil temperature values associated with all N2O measurements. If not, how was this approached?

AR: Thank you for raising these important points. The number of observations for each phase was calculated based on the number of sampling dates within that phase. While N₂O fluxes were measured weekly, soil samples for SMN and WFPS were collected biweekly. To align these variables with the weekly N₂O flux measurements, we performed linear interpolation for SMN and WFPS values between the biweekly sampling points. This approach ensured that all N₂O measurements had associated SMN, WFPS, and soil temperature values for the modeling. We have revised the methodology to clarify this process.

RC2: Lines 389-390: no need to re-state how the calculation was conducted.

AR: Thank you for pointing this out. We have removed the redundant sentence in the revised manuscript.

RC2: Lines 435 to 444: authors discuss the impacts of cover crops on soil properties that contribute to larger soil water holding capacity. However, this is likely not the explanation for the results observed in the present study (no differences between treatments), as cover crops were implemented only for a single cropping season in each of the sites. Consider narrowing your discussion and emphasizing aspects that may further explain what was observed in your study.

AR: Thank you for this suggestion we have narrowed the discussion regarding the effect of cover crops on WFPS.

RC2: Lines 481-482: this other paper reports similar findings: https://www.sciencedirect.com/science/article/abs/pii/S0167880921004540

AR: Thank you for the suggestion. This reference has been added to the revised manuscript.

RC2: Lines 509-511: consider adding a citation to this sentence.

AR:  thank you for the suggestion. We added a citation to the revised manuscript (Signor and Cerri, 2013).

RC2: Lines 530-535: what about the impact of freeze-thaw periods on N2O emissions? You mention that soil temperature dynamics affect N2O emissions (line 535), but this is not discussed in the specific context of your study. For example, could plots of frost-sensitive cover crops exhibit more freeze-thaw cycles due to less biomass insulation of the soil, and therefore more N2O emissions? This point is discussed to some degree later in the section. Consider expanding the discussion, and potentially merging paragraphs addressing this topic so overall flow is improved.

AR: Thank you for raising this important point. We believe that we did not explicitly capture N₂O fluxes from freeze-thaw cycles in our study. Specifically, elevated N₂O emissions following frost events were predominantly observed in frost-sensitive cover crop treatments, while the frost-tolerant treatment and bare fallow showed no similar increase. This observation suggests that the emissions are largely attributable to the physical damage inflicted on frost-sensitive biomass by freezing temperatures and not the frost-thaw of the soil. This damage likely releases readily available carbon (C) and nitrogen (N) substrates, which can fuel microbial activity and enhance N₂O production during thawing periods.

RC2: Line 536-541: these sentences do not seem to fit very well together, as authors discussed fertilizer application, organic amendment application and tillage practices in a few sentences; consider rephrasing.

AR: We appreciate this suggestion. We have restructured the sentences to enhance their clarity and coherence.

RC2: Line 592: Authors mention “frost-induced N2O emissions”; how were those emissions defined or calculated? Did you consider variations in soil temperature above and below zero during the time of the year in which freeze-thaw emissions are likely to occur?

AR: A possible way to define these emissions would be to sum all fluxes following sub-zero soil temperatures, identifying the specific periods when freeze-thaw cycles occur. However, in our study, we opted not to quantify frost-induced N₂O emissions. We utilized the closed chamber method, which involves periodic measurements of N₂O fluxes. While effective for many purposes, this method may miss rapid flux peaks associated with freeze-thaw events, which are often short-lived but significant contributors to cumulative N₂O emissions. Consequently, our study may have underestimated the total N₂O emissions during the winter period. To address this limitation, we have mentioned in the discussion the advantages of continuous measurement systems, which would better capture such transient emissions. This point underscores the need for future research to include high-resolution temporal measurements during periods of freeze-thaw dynamics to gain a more accurate understanding of their impact on N₂O emissions.

RC2: Lines 627-633: great to see these lines discussing the limitations of the potential impact of cover crops on mitigating indirect N2O emissions. I also wonder how this mitigation potential would look like when considering an LCA approach for the entire crop rotation you studied. If the N in the cover crop biomass mineralizes again during the sugar beet and/or winter wheat phase, could we still affirm these indirect emissions were mitigated? Consider discussing this point. Related to the previous point, I wonder if it would be possible to incorporate what the net GHG emissions in CO2-eq for each treatment would look like, when considering N2O emissions, C sequestration, and potentially avoided indirect N2O emissions (although not sure if I would bring in this last piece, per item mentioned above). Consider incorporating this into your results and discussion.

AR: Thank you for these thoughtful points. We have addressed them as follows:

1. Differentiation between direct and indirect N₂O emissions:

    We agree that it is essential to distinguish between direct and indirect N₂O emissions in the context of our study. Even if the N in the cover crop biomass mineralizes during subsequent cropping phases, leading to increased direct N₂O emissions (as observed in the rye treatment); the mitigated indirect N₂O emissions would remain unaffected. However, the net N₂O budget for the system would indeed be influenced. In the discussion (section 4.5.1), we stated that the use of cover crops has led to slightly higher N₂O emissions, which can be mitigated to a certain extent by the reduction of indirect N₂O emissions and enhancing C sequestration through the cover crop cultivation.

1. Net GHG emissions in CO_2-eq:

    We appreciate the suggestion to incorporate an estimate of net GHG emissions in CO_2-eq for each treatment. However, this is beyond the scope of our current study due to the lack of critical parameters required for such calculations. Such as the CO₂ emissions associated with the production, sowing, and management of cover crops, NH₃ emissions and energy use and associated emissions from machinery for tillage, residue incorporation, and other field operations that are associated with cover crop cultivation. Incorporating these components was not within the framework of our study. However, we acknowledge the importance of this perspective and suggest it as a critical area for future research to evaluate the full mitigation potential of cover crops within crop rotations. We will highlight in the discussion section that future studies should adopt an LCA approach to assess the net GHG balance. This broader assessment would provide more conclusive insights into the sustainability and mitigation potential of cover cropping systems.