This study is about the root zone storage capacity Sumax in the Upper Neckar river basin in Germany.

Three null hypotheses are:

(1) Sumax significantly changes over multiple decades reflecting vegetation adaptation to climatic variability,

(2) changes in Sumax affect the long-term partitioning of drainage and evaporation and thus control deviations ΔIE from the catchment-specific trajectory in the Budyko space

(3) a time-dynamic implementation of Sumax improves the representation of streamflow in a hydrological model.

I enjoyed reading this manuscript, although it took time. The manuscript is well-written and organized effectively. The topic is insightful and stimulates new thinking about watershed hydrological dynamics. However, I have a few comments to offer. These comments are intended for open discussion and do not critique the methodology of the Sumax or the manuscript.

Major comments:

1. The abstract is unnecessarily long and contains repetitive statements. I suggest making it more concise.
2. I am confused by the use of the terminologies "transpiration," "evaporation," and "transpiration" in certain contexts. For example, in Lines 198-205, Equation 5 should represent the overall water balance in a watershed, thus Er_bar should indicate total evapotranspiration rather than just transpiration. In Line 203, author used “potential evaporation”, “evaporation”. These make me lost.
3. Somehow, I am unable to access the data and model in the Code and Data Availability section. Although HESS does not mandate the openness of data/code/user-guides like the Geoscientific Model Development does, I encourage the author to make these accessible to enable readers to replicate or advance the work, thus expanding its impact.
4. I question the solid physical meaning of the Sumax. Firstly, Sumax is not a directly measurable feature using devices; it seems to be derived from known variables (precipitation, ET, streamflow). Such derivations generally should have a clear meaning, indicating their driving factors. Hence, the question arises: what are the driving factors determining the value of Sumax? Can it be measured without long-term climate data or model calibration? The equations 1-8 calculate Srd.n rather than derive Sumax. The concept of Sumax seems more akin to a feature in a conceptual model, derived from data. Unlike field or laboratory measurable parameters like conductivity in soil flux calculations via Darcy’s Law, Sumax cannot be directly measured or validated experimentally.
5. Sumax is derived from the differences between effective precipitation and transpiration. The calibrated values of Sumax (Sumax,cal) are computed using outputs from the FLEX model, which is calibrated by streamflow. This suggests that the streamflow simulations are reliable within the FLEX model but may not imply the reliability of ET calculations (PET, AET, evaporation, transpiration, etc.). I would like to hear your thoughts on this challenge.
6. If Sumax can be derived from observed precipitation, ET, etc., what is the necessity for model calibration? Consider a hypothetical experiment: if someone sets the Sumax value in the FLEX model based on observed data and then calibrates the model using streamflow, would this experiment yield comparable performance metrics (e.g., NSE) to those obtained from the model/simulation? I know you already test the model output via fixed Sumax, but did not focus on the NSE performances.
7. An opinion paper by Gao et al. (2023) (10.5194/hess-27-2607-2023) discusses concepts that may connect to the soil features or arguments presented in this manuscript. I am neutral on the opinions expressed in Gao et al. (2023), but I am curious whether there are links between the Sumax concept and the points made in this paper.
8. The paper attempts to establish a connection between Sumax and vegetation adaptation to climate. However, I do not see any analysis on vegetation adaptation, except for the use of omega in the Budyko method. Moreover, the ET is an output from the model, not an observation linked to vegetation-specific features. Given these uncertainties, I believe the current findings are sufficient for publication and recommend not expanding them to include vegetation adaptation.
9. I encourage the author to disclose all the calibrated parameter values from the model. These values indicate both the performance of the model and the characteristics of the watersheds.
10. Figure S1: The groundwater storage (Ss) in the figure implies a seasonal variation. What factors cause the seasonal variation of groundwater storage? The variation of groundwater storage implies the variation of baseflow, but it did not affect the total streamflow (Q). Is there any data/analysis support the groundwater storage and baseflow? There are two more reservoirs (unsaturated fast) in the model. Could you show the outputs about them?
11. Let's conceptualize an ideal watershed based on your presented data. When the maximum water deficit in the root zone (Sumax) is about 120 mm (your results), and the maximum groundwater storage is approximately 4 mm (Figure S1), assuming typical porosities for the two layers (p_root = 0.4 and p_gw = 0.2), the calculated depth of the average hydrological-response aquifer would be 120/0.4+4/0.2=320 mm. This value represents the aquifer depth necessary for the hydrological response in this watershed. However, this formulation does not account for the unsaturated and fast reservoirs, as these are not detailed in the manuscript. I wish to highlight two concerns: (1) the calculated 320 mm depth for the hydrological-response aquifer seems unreasonable; (2) there is a need for information and analysis concerning the unsaturated and faster reservoirs to better understand the watershed dynamics and model structure.

Minor comments:

1. What is the f(x) in equation 7?
2. Line205: what is the El-bar in your equation?
3. Line219-221: citation is necessary.
4. Line790: I don’t see the figure2(b) make any sense here. The accumulation of ET in such long period does not tell clear message here. The shaded areas for the t1-t4 are not very clear in these figures. I am not sure, but the maximum streamflow in figure 2d seems beyond of the y-axis-max and was crop out of the figure box.
5. Table S3: I interpreted the values in the table, for example, “0.59(0.06-0.55)” in the first cell, as “mean/media (min – max)”, but the value “mean” is out of range of min-max. Do I misunderstand the meanings of the value in the table?
6. Line348: No section 4.1.2 in this manuscript.
7. Line 352: What is the p value here? They seem not the common p-values in statistics. I don’t think the 4-sample analysis can tell any potential relationship between the two variables, let alone any convincing conclusions.
8. Line 471-472: “The catchment-scale root zone storage capacity (Sumax) is a critical factor affecting the moisture exchange between land and atmosphere as well as the hydrological response in terrestrial hydrological systems”. The “affect” may not the right word, “reflect” may be.
9. Figure S1: I cannot find the gray shades in the figure. Or are they fully overlapped with green shades?

This manuscript conducted a systematic analysis on the temporal evolution of the catchment-scale root zone storage capacity (Sumax) and its influence on hydrological response. Two methods were adopted to determine Sumax separately, ensuring a credible estimation on Sumax. The influence of changing Sumax on hydrological processes was analyzed in two aspects, focusing on the drainage/evaporative flux partitioning and the short term hydrological response dynamics. Overall, the logic and key points of this paper are clear, and the analysis is solid. However, I still see some improvement room. I agree with the major and minor concerns raised by the reviewer #1, and I would like to offer four additional major comments.

1. The authors divided the whole period into four subperiod to calculate the Sumax, its relation with climatic indices, and its influence on hydrological response. However, I would question whether such a division could produce reasonable results. First, many climatic indices don’t show significant difference among four periods, making it difficult to see the relation between these indices and the Sumax. Second, regression based on only four points has large uncertainty and occasionality. For example, in Figure 10, if we remove the point with largest Sumax, a significant negative relation between ΔIE and Sumax can be obtained. Maybe the authors can attempt to increase the number of subperiod or discuss this issue in a limitation section.

1. Although the overall logic and main conclusion of this manuscript are clear, some of the methods and results are presented too complicatedly. Some specific suggestions from my reading experience:

• For most of figures, I cannot see the necessity of using gradual color to distinct the results of different period, since they can be clearly distanced by the x-axis. Instead, for Figure 9b, I think showing the period of each point by different color would be better.
• There are lots of variables in this paper. I would like to suggest the authors to provide a table to show the meanings of all the variables to make the paper easier to follow. Besides, if I don’t miss something, I think some variables are not explained. (1) Equation 7 is confusing. What does f(x) mean? What does Srd(t) and what is the difference from Srd,n(t)? The meaning of n is not explained. (2) The subscripts o and o’ described in 5.4 haven’t appeared in the method section. I guess it may be explained in the missing 4.1.2 section.
• As pointed out by another reviewer, the abstract is too long. The three paragraphs are actually telling one thing, that is, the three hypotheses and the related to them. I also suggest the authors to change the expression of the hypotheses to the form of scientific question, at least for the first paragraph of abstract. I was really confused when I read the second hypothesis for the first time because it was contradictory to the title, and finally I realize that it is just a hypothesis which is rejected later. I think express them more straightly could help readers get your main conclusions more easily.

1. For the Sumax determined by hydrological model, the authors regarded all parameters on the pareto front as feasible. However, there are some extremely low values for some metrics such as NSEQ and NSElogQ. I think it would be better to select the behavioral solutions based on the threshold of each metric for analysis. Also, I would like to suggest the authors to present the metrics for each subperiod produced by scenario 1, and that for the whole period T produced by scenario 2 in Table S3, to allow for a direct comparison between two scenarios.

1. Although the calculation and analysis are solid, the main conclusion of this paper is not so favorable for its publication. The results indicate that the change of Sumax neither controls the drainage/evaporate water flux partitioning, nor affects short term hydrological response dynamics, and considering the variation of Sumax also leads to little improvement in hydrological model performance. So a reader may question why we need to care about Sumax. I suggest the authors to add some open discussion on the significance of Sumax and its influence on hydrological cycle. Besides, given that the conclusion is different to some other studies, it is strongly recommended to discuss what factors determine whether the hypotheses 2 and 3 would be rejected, i.e., in what kind of catchments, considering the change of Sumax would improve model performance? This will make the conclusion of this paper more general and useful.