This manuscript presents a numerical modeling study exploring how inter- and intra-annual precipitation variability affects nitrogen (N) loads and fluxes in a catchment. The topic is timely and important for water quality and environmental management. The paper is generally well written, though some grammatical edits are needed. The methods and results are mostly clear and logically presented. The conclusions are relevant and likely to be of interest to scientists and resource managers focused on mitigating nutrient pollution.

Primary suggestions are to:

(1) Clarify and narrow the research focus: While the broad importance of precipitation variability is well established in the introduction, the final introductory paragraph should better define the unique contribution this work makes to the literature. Consider making the statement of objectives more specific and clarifying how the objectives fill gaps left by other recent studies.

(2) Differentiate from previous work: The study references and builds on previous work, particularly Wang et al. (2023). However, the manuscript does not always make it clear where prior work ends and the current study begins. The manuscript also refers the reader to the previous studies for some crucial details (e.g. calibration to stream nitrate concentrations), which are not easy to find in the previous studies. I recommend adding further descriptions of the related studies at this site and delineating which analyses, model developments, and findings are new in this study.

(3) Clarify model-data connection: A clearer connection could be made between the numerical model and real-world observations. The use of simple abstractions in the nutrient transport component of the model is reasonable, but also requires careful consideration of (a) how well the simplified processes representations mimic actual processes, (b) the uncertainty of the parameter estimates, and (c) the accuracy of the model in terms of reproducing observations. Otherwise, there is a risk of circularity: the model is built around certain processes and parameterizations, and then used to test the importance of those same processes and parameters.

Line-by-Line comments:

Line 14: “performance” is wrong word. “effect”?

34: SON not defined at this point

39 – “a small effect”

58 – not clear what “their” refers to

59 – grammar problem

72 –“a major”

59-78 – the paragraph starts with climate variability and ends with nitrate. I recommend keeping to one topic per paragraph.

99-100 sentence fragment

105 – “it is”

107 – The objectives are somewhat broad. It has been established that precipitation variability can affect N dynamics, and it would help to be more specific in this paragraph about the aspects of variability being tested and what if anything has been done to address them previously. In other words, how the specific objectives of this study relate to gaps in knowledge left by previous studies?

163: cross section is not discernable and it is not clear what is the source of the saturation values

164: this figure is mostly recycled from Wang et al., 2023 but no citation is given.

169: Please clarify if/how these data are used in the current study.

192 – delete “in details”

231 – what does this mean that the calibrated model was “verified” over the entire simulation period?

240 – meaning unclear “delineated corresponding to the reality”

245-6: “in route” grammar

258-260: meaning unclear.

320: “validation” might not be the right word (not the same as calibration).

323: Wang et al 2023 refers readers to Yang 2018 for more details and is not an easy source of information about the estimation of the N cycling parameters, uncertainty of those parameters or the quality of fit to the data. These are crucial aspects of the calibrated model and should be presented clearly and succinctly for the readers.

325: “impermeable for nitrate” (and water?)

333-375: probably don’t need this much detail about the rainfall generator

336: “a stochastic model”

Table 1 – This table has too many numbers and variables for readers to easily absorb. Consider replacing with a schematic, examples, or another simpler figure or table.

399-401 – grammar problem, meaning is lost

458 – Why is soil denitrification lumped with GW denitrification? Are they expected to be similar?

503 – These figures are confusing because the response variables (SON, SIN, LEA, Cq) are not on the z-axes.

564: 5.1 section title: consider being more specific about what increased rainfall does to the N dynamics

608: It seems notable that the high flows during the 2018 to 2019 drought are as high as the high flows from 2014-2018, and the main difference during 2018-2019 seems to be in the low flow periods.

670: 5.6. Consider discussing: Data limitations, uncertainty of parameters, model process representations

Table S1 – van Meter reference is missing date; bibliography is not included in this document

The authors present an original piece of research focusing on the catchment response of N fluxes to rainfall interannuel and intra-annual variability using synthetic experiments based on the Hydrogeosphere model. 

I found the study clean and rigorously described, calibration method is sensible. Figures provide useful and clear illustrations of the results. The manuscript is well written. I think the discussion could be expanded and therefore I recommand a minor revision of the manuscript. 

Especially I would have 2 comments on the discussion: 

1) several parts of the discussion present some reactions as simulation results when they are a direct consequence of the modelling equations. Ex.: Lines 566 to 568, the impact of temperature and wetness on mineralization is constructed mathematically in equations 1 and 2, isn't it? Same comment regarding plant uptake and denitrification. According to my opinion, the interest of the model is rather to calculate which of these mechanisms is going to dominate the response, and also it helps to consider different time scales of response, which is particularly relevant for droughts (cf. lines 97-99)  

2) I have a general comment on the modelling choice that is N mechanisms are much more simplified than the water processes. While I am very aware of the computationnal challenges associated with such virtual experiments, I find intellectually disturbing to have a fully mechanistic approach to represent water combined with a representation of nitrogen very simplified in comparison. What do you think?  There is no representation of plant-soil processes, and considering the results of the EDY scenario in terms of N dynamics, the response of plants growth or death to water stress seems to be a key mechanism. Also, about the fact that all external nitrogen inputs are introduced in the SON pool (lines 250, 252), I was wondering to what extend it refers to a reality?  Are fertilizers mainly applied as urea? Also, I did not understand why is this framework acceptable and according to which criteria or arguments (line 258-260)? 

I have also a few minor comments below: 

3) line 41: I suspect that fertilization strategies  will evolve with the changes in timing of plant growth stages with warming 

4) introduction should maybe be restricted to context elements that are directly linked to the study, lines 45 to 61 : it could go more directly to the point of rainfall variability skipping the details on extremes. Lines 79 to 82: I missed the link with the present study here. 

5) lines 145-146 : at which frequency are concentration measured? 

6) line 339-340 what is a "good perfomance under Climate change" for the stochastic generator? 

7) Figure 4 a: it would be useful to have measured C_Q in the plot too. The legend refers to "acceptable simulations" but so far as I understand it is more the variability associated to the generator in each scenarios (n=100) isn't it? 

8) lines 577-578 : are preferential flowpaths represented in the model? 

9) it would be useful to see the effect on water fluxes as well: especially recharge flux for groundwater and Actual ET in Figure 10 or Figure S2 (lines 653 to 656)