The Santos et al. manuscript presents an interesting study on the integrated water resource management (IWRM) modelling approach to explore the impacts of climate change and future demand scenarios at a catchment in Northwest France.  While the overall scientific approach appears to be sound, I have a few comments and suggestions below, which will hopefully improve some aspects of the paper's presentation:

1) The Methods section presents three versions of the hydrological model: calib, uninf, and iwrm.  If I understood it correctly, only the calib version of the model is used for parameter calibration and model validation using the observed streamflow data, and the other model variants, uninf and iwrm, do not have the corresponding observed data to gauge their performance.  However, Table 5 in the Results section shows the model performance of all three variants.  So, what are the uninf and iwrm variants being compared against?  And what is the point of showing which model variant performs best?  As I understand it, each model variant serves a totally different purpose, and they are not competing against each other.

2) In my opinion, the water demand and release models of the iwrm variant are the most important contribution of this paper.  However, more information might be needed to determine the robustness of the water demand and release models presented here.  Not much information has been provided about the input data used for the models described in Appendix B.  Where has this data been sourced from?  Is the model implementation done in spatial grids, or is it spatially lumped at the subcatchment level?  The cattle watering model equation allows for information on different cattle types.  How many different cattle types were considered?  And how is their water demand calculated?  The same question applies to the demand calculation for other uses.  Also, why does the formula for drinking water demand add cattle and industrial water demands to the population's consumption?  Perhaps it might be useful to provide a detailed schematic, maybe at a subcatchment level, of how the different water demand models are feeding into the base hydrological model to create the iwrm variant.

3) Another potentially innovative aspect that has unfortunately been sidelined in the paper is information from the stakeholder workshops.  I think more detailed information is needed on how the three future scenarios were initially designed and on the specific value added by the stakeholder feedback.  As currently presented, we are only seeing the final product, and the importance of stakeholders in shaping these scenarios for the local conditions is being ignored.

4) Lastly, while the iwrm model presented in this study seems innovative, it is certainly not the first one to have attempted a quantification of future water demand.  There have been a large number of studies conducted using other models, most notably WEAP, to address water demand management and forecasting.  In this context, I find it troubling that the presented iwrm model, and its results, have not been discussed in the context of other existing models.  It would be quite valuable for the authors to discuss the similarities and differences in the specific aspects of their iwrm model and others found in the literature.

Review of “Exploring future water resources and uses considering water demand scenarios and climate change for the French Sèvre Nantaise basin” submitted to Hydrology and Earth System Science for consideration for publication. 

The manuscript describes modelling study on the impact of climate change and water demand/supply scenarios on water availability in a French catchment. The authors analysed the current water supply and demand, based on existing data. A hydrological model was set-up and calibrated, considering water use in several sectors. The model was subsequently applied to a climate change scenario, considering 5 climate models. Moreover, 3 water use scenarios were defined and their impact assessed. The study shows that climate change will affect the water balance in the catchment, but that future water use does not differentiate that much between the different scenarios. 

Overall, the manuscript is well-written and study is well-performed. However, there are some revisions needed. Especially the presentation of the results should be improved. It sometimes seems that I’m reading a data report, rather than a scientific article. The authors show so much results that is difficult, if not impossible, to extract the main message from each figure and table. Moreover, improvements to the Introduction (structure), Material & Methods (too much detail) and Discussion (too little detail) are needed.

The Introduction needs some revision. The most relevant concepts and previous research on the subject are described. However, the structure is not that clear. The authors suggest 2 options for studying water supply and demand in catchments. It is not clear what is actually the difference between the two options. Moreover, which of the two options is used in this study? If the two options are very relevant for the study, then this needs to be more integrated into the rest of the Introduction. For instance, the sentences prior to the objective should focus on these two options and the two options should be integrated into the objective as well. 

The Material & Methods section is rather long and provides maybe too much detail. Lines 139-143 and 194-199 give information that seems not too relevant for the study. There are also a few paragraphs that can be shortened without losing much relevant information, such as those paragraphs in lines 200-213. Many paragraphs start with a few (2-3) introductory sentences that can easily be replaced with a single sentence (e.g. lines 267-269 can be replaced by “Management rules were implemented in case of water deficit within the system.”) Finally, I also suggest to move some figures and tables to the Supporting Material, such as Figure 2 and Tables 2 and 4. 

The authors use an ensemble of 5 climate models (GCM-RCM combinations) which is very important to account for climate model uncertainty. The authors show the results of each climate model individually, which leads to rather overwhelming figures with a lot of data, especially Figures 5-9. I highly suggest to show the ensemble average instead and determine the uncertainty of the results using an uncertainty band or use statistics to quantify uncertainty. This would make the figures much more interesting to look at. At the end, the readers will be interested in the general tendency of the results. The same holds for the results show for each individual management unit. This becomes much more appealing to look at if you show the results on a map. Readers are not going back and forth between the study area map and the results to see where runoff is going to increase or decrease. Readers are also not going to compare the results of all individual climate models. So, please reduce the amount of data shown in all these figures and try to provide enough data to make a coherent story. 

The Discussion section needs to be significantly improved. Most discussion stays on the surface and does not go deep into the literature on how the results if this study compares to previous results. Neither do the concepts the authors refer to in the Discussion are backed by previous research. There is a lot of room for improvements in the Discussion.

The same can be said for the Conclusions. The Conclusions should be one-on-one related to the objectives of the study. I don’t have the feeling that the Conclusions answer the objectives/research questions posed in the Introduction. Currently, it reads more like a summary, but includes also several lines about the implications of the study. So did the authors reach the objective of the study? And to what extent? Such questions should be answered in the Conclusions.

Below I have provided specific comments to the text, figures and tables.

Specific comments

Lines 19-22: The authors refer here to water demand satisfaction, but it is unclear what is meant by that. Can this be quantified? I guess this must be the case, because the authors use a water resources model in their study. It would be useful if the authors can give quantified estimates of the water demand satisfaction or give a clear definition what is meant by it.

Lines 19-20: Please clarify in the text what you mean with “A single climate projection” and “a less drastic deterioration of the system”. 

Lines 41-47: It is unclear what the main message is of this paragraph. It seems somewhat directed to the fact that hydrology can be studied with models. I guess this can be included in a single sentence to start the next paragraph.

Lines 51-52: You mean water supply and water demand studies? Please clarify in the text.

Line 51: “This approach...with demand”. This part of the sentence largely repeats the previous sentence. Please remove or revise.

Lines 52-53: Are hydrological models really needed to estimate water supply and water demand? From the previous sentences it does not become clear why that is the case. Please clarify.

Line 55: What is meant by coarse models? Please clarify in the text.

Lines 56-60: It is difficult to understand the difference between the two approaches, because both seem to involve models. So what is the actual difference? Please clarify this in the text.

Lines 67-68: It seems from this sentence that integrated and sustainable water resources management is a concern, but I suspect that the lack of management will be a concern, not the management itself. Please revise this sentence.

Lines 76-77: This refers to the first option from lines 50-55? In that case, say first option or give this option a specific name in lines 50-55 and refer to that name. This would clarify this for the readers.

Lines 67-81: These sentences put the rest of the Introduction into a local perspective, which is fine. However, there is quite some repetition in these sentences and this local focus is maybe better suited to go into detail under Material & Methods and Study Area. I suggest to shorten this part of the Introduction to a few sentences in which the main points of the Introduction are summarized from which follows the objective of the study. So what is the research gap the authors are focusing on? 

Lines 85-88: These specific objectives do not add much to the two main objectives. I suggest to replace these sentences with a few (2-4) sentences in which the methods are briefly explained. 

Line 94: Replace “who” with “which”.

Figure 1: The legend says that with blue the rivers and dams are indicated, but it is in fact the rivers and the reservoirs. It would be useful to indicate the dams as well. I’m not sure if it is necessary to indicate the management units and the sub-catchments. The location of the sub-catchments might be arbitrary or is there a physical reason for their location?

Lines 108-109: What is considered upstream and downstream? Please clarify in the text.

Table 1: I suppose this table does not contain the data for the large reservoirs. Then there must be an upper limit for the capacity of the third category. Please indicate this in the table.

Lines 166-168: Why did the authors consider the RCP8.5 scenario? Please clarify in the text.

Lines 168: Please explain briefly how the ADAMONT bias-correction method works.

Lines 168-172: The two criteria used to select the 5 GCM/RCM combinations should be better described. How did you compare the EURO-CORDEX projections with the CMIP6 projections? What do you consider representative with respect to the RCP8.5 scenario at the end of the 21st century? Please clarify these issues in the text.

Line 179: Replace “to an increase” with “in an increase”.

Table 3: I suggest to replace the symbols (+, -, =) with the actual change in percentage for precipitation and degrees for temperature. Or otherwise, explain what the symbols mean. But my preference goes to quantified estimates of the change in precipitation and temperature.

Line 191: With “heads” the authors mean “animals”?

Lines 191-193: Please indicate how many years data are available for irrigation and drinking water withdrawals. It seems a bit unclear now how much data are available.

Line 227: Please revise this sentence and explain the parameters in a logical order, i.e. “X1 is the production capacity parameter (mm), X2 is the inter-catchment exchange coefficient (mm d-1), X3 is…”.

Lines 235-237: I guess each sub-catchment is considered a homogeneous unit within the model. How does land use and soil characteristics affect the behaviour of the model? Should each sub-catchment not be homogeneous considering land use and soil? Please explain in the text.

Lines 244-246: So the authors applied the GR6J model using the airGR R package? Please indicate this in the previous subsection.

Lines 269-272: Can the levels of restriction be quantified? Are they related to the discharge or total water availability? Please clarify in the text.

Line 270: With “interdiction” the authors mean “suppression”?

Line 306-308: This was indicated already a few lines back, please remove.

Line 335: Why is this a remarkable result? Please clarify in the text.

Line 346: I’m not familiar with using the correlation (between observed and simulations?) as a model performance metric in hydrological modelling studies. Given the low impact on the results, I suggest to remove this from the manuscript.

Lines 363-370: It is perfectly fine to remove this paragraph and first sentence of the next paragraph, which mostly repeats what is already included in the Material & Methods section. 

Line 368: What is meant by MU scale?

Figure 5: Here the authors show the results of 2 future climate projections that were not mentioned in the Material & Methods and are neither included in all other figures. I suggest to remove these two projections and solely focus on 2056-2085, this would make the manuscript much clearer to follow.

401: Please delete this sentence, it repeats the title of this subsection.

Line 410-412: It seems that irrigation water demand is increasing in all three scenarios, also for the alternative scenario. So why do the authors claim that the alternative scenario leads to an overall decrease? Please clarify in the text.

Lines 412-414: I don’t see much differences between the MUs. Like argued before, try to simplify the results and show general tendencies, especially when there are no differences between MUs and between climate models. 

Figure 7: The order of the scenario is consistent with the description of the scenarios in lines 326-327, i.e. constant, trend, alternative. Please adjust the order according to the Material & Methods.

421: Please delete this sentence, it repeats the title of this subsection.

Lines 429-438: These sentences are very hard to follow because of the focus simultaneously on MUs and scenarios. Please revise.

Figure 8: I’m really lost at this point. The terms “influenced indicator” and “uninfluenced indicator” only appear in this caption. There is an uninfluenced model, does this refer to the uninfluenced indicator? I had to look up that QA refers to the mean streamflow, why would an abbreviation be needed here? The same holds for the low flow indicator. I suggest to use mean flow and low flow instead of QA and QMNA5. Apart from that. Is it not more logical to present these results as a change from the reference scenario (uninfluenced?) to the scenarios with climate change (influenced?)? The mean flow seems not to change, that seems odd. 

Lines 473-484: The first part of this paragraph focusses on the climate change results, while the second part focuses on the scenarios. I suggest to separate these two discussions.

Lines 475-478: There is no need to repeat these details in the Discussion section. Please focus on the discussion of the results. Please remove.

Lines 478-479: Most readers are unfamiliar with the Explore2 project. So in what sense are the results coherent with this project? Please clarify in the text.

Lines 485-508: I’m not sure what point the authors try to make here. So the scenarios do not show much differences between each other, apart for certain climate models. The latter should not be included in the discussion, because of the high uncertainty in the climate model output. Again, the average climate model ensemble should be considered, instead of the individual climate models. Regarding the scenarios. The information provided in the appendix that shows the differences between the scenarios does not allow to compare the different scenarios so easily with each other. It would be better to present 1 table in which all three scenarios are shown next to each other. So I cannot judge if the scenarios are that different from each other. It seems from lines 496-498 that the authors already expected that the scenarios were not so different from each other, so why not focus on only 1 scenario? How did other studies deal with this?

Lines 509-532: This subsection gives very little discussion on what I would expect from the title. I was expecting more in-depth discussion on the model structure and how to incorporate water resources data into these models. The presented discussion stays very much on the surface, without going deeper into the literature on how previous studies dealt with these issues. Please revise.

Lines 534-555: Obviously, the authors should assess the climate model uncertainty. Similar to the previous subsection, this discussion stays on the surface. It highlights the uncertainties that are at play, but does not discuss which uncertainties are more important in the current study and how that compares to previous research. Please revise.

Lines 557-562: Please reduce the description of the methods to a single sentence and focus on the conclusions of the study.

Lines 562-563: Please remove this sentence.

Lines 567-568: I’m not sure to which scales the authors are referring to in this sentence. What is the catchment-scale in this case? The whole catchment or the MUs? What would be a larger scale? The authors did not perform analysis at larger scales, so how can they be so sure that at larger scales these processes are not becoming visible? Please clarify in the text.

Lines 569-575: These are more implications of the study and go beyond the current study. This would be better fit for the Discussion section.

GENERAL COMMENTS

The paper aims to characterize the future water availability and demands in the Sèvre Nantaise basin (France), using a set of climate projections and future demand scenarios as inputs for an integrated water resources model. According to the results, low flows and water demand satisfaction will decline in the future, mainly driven by climate change.

From my point of view, the proposed methodology lacks novelty as it does not include new concepts, ideas or methods but it may provide useful insights for future management in the Sèvre Nantaise basin. The paper is generally well written, although it is too long and sometimes difficult to follow. Besides, some aspects require further clarification.

SPECIFIC COMMENTS

• Abstract: please mention that the modelling approach was applied to the Sèvre Nantaise basin.
• Introduction (lines 62-64): “To anticipate how both resources and uses will evolve, and how conflicts may arise due to water shortages, these models need to be applied in conjunction with climate projections (Hadri et al., 2022) and water demand scenarios. In my opinion, the introduction should further discuss the limitations of approaches based exclusively on climate projections and the existence of methodological frameworks aimed at characterizing the system’s response over a wide range of plausible futures instead (e.g. Decision Scaling, Scenario-Neutral approach, Info-Gap theory).
• Introduction (lines 67-81): I suggest summarizing this paragraph, as it provides context on the particular case study but does not include further scientific perspectives.
• Figure 1: Please make this figure more informative (e.g. other demand nodes apart from urban areas). A scheme of the water resource management model will be useful, including all the nodes considered for the analysis.
• Lines 109-110: “Based on the observed data from the Nantes meteorological station, located close to the catchment outlet, the temperature increase is about 0.2 to 0.3 °C per decade”. Is there any spatial heterogeneity over the basin with regard to this increase?
• Lines 121-122: “The hydrology of the Sèvre Nantaise is influenced by several hydraulic structures designed to meet the water demands of various human activities, including irrigation, drinking water supply, cattle watering and industry”. Please indicate the share of the total annual demand for each use. Does rainfed agriculture constitute the predominant land use? As agricultural areas cover 88.2% of the basin (line 97) but irrigation demands are less than drinking water ones in some MUs (Table 4). What are the existing types of crops?
• Lines 123-124: “The Bultière reservoir, located on the main course of the Grande Maine River (Figure 1), has a maximum capacity of 5 10⁶ m³ and supplies water to a wastewater treatment plant with a maximum daily production capacity of 22,000 m³”. I understand that the reservoir supplies water to a water treatment plant (for water purification and subsequent user supply).
• Lines 124-125: “The Ribou-Verdon reservoirs, located on the Moine River, have a combined maximum capacity of 17.6 10⁶ m³ and provide drinking water to a treatment plant”. I understand that it means the provision of raw water to the treatment plant instead.
• Lines 131-132: “In addition, water transfers between neighbouring catchments are possible”. How much water is transferred from/to other basins?
• Line 147: “Climate data from the SAFRAN atmospheric reanalysis were used”. Please clearly indicate for which period.
• Lines 155-156: “It can be seen in Table 2 that the ratio between Q25 and Q75 can differ a lot between very contrasted stations (e.g. on the Sanguèze) or less contrasted stations (e.g. on the downstream Sèvre Nantaise)”. Does this mean that some of the stations are less reliable? How has this been considered for the model calibration? I suggest moving Table 2 to the Supplementary material.
• Line 168: “using the ADAMONT bias-correction method”. Which period was considered for the bias correction? 1976-2005? In my opinion, it would be interesting to see the differences between the SAFRAN reanalysis and the GCM/RCM outputs during the baseline period (2008-2020) as well.
• Lines 205-206: “wastewater treatment plants exhibits greater variability. This variability is likely attributable to fluctuating precipitation levels that locally contribute to the release”. This explains why releases from the water treatment plants are generally higher (except for a couple of MUs) than the drinking water supply demand in Table 4. Depending on the type of wastewater network, this modifies the point of the system in which the drainage water joins the stream and the timing (e.g. storm tanks). Hence, I understand that the runoff generated in urban drainage areas is excluded from the surface runoff to be accounted as part of the treated wastewater release instead, is that right?
• Table 4: what about returns from agricultural areas, the main land use in the basin? Are the water transfers to other basins accounted as part of the drinking water demand? It will be useful to reflect the intra-annual variability of the water demand.
• Lines 268-270: “when the water resources are insufficient to satisfy all the uses simultaneously, which is estimated through the use of streamflow thresholds, the model prescribes restrictions with different levels”. Are these thresholds in accordance with an existing drought management plan in the basin? Please cite it if this is the case.
• Lines 272-274: “In circumstances where streamflow is at its lowest, the model implements an order of priority for the uses, emulating the real-world practices, as follows: drinking water production, cattle watering, industrial production, irrigation, and finally, small reservoirs filling up”. What about environmental demands? Are they consider as a model constraint to ensure that the environmental flow is preserved?
• Exploratory scenarios for projections of water demands (line 314): How do these future scenarios relate to the SSPs? Why is it considered that irrigation practices will not change under future scenarios (Table 1, Appendix C), given expected temperature rise and precipitation decrease? Neither the irrigated area?
• Lines 341-342: “Looking at the bias score, we observe that streamflow is underestimated when the dam is not neglected or represented by a model”. Is there a problem with the defined operating rules for the dam? Please discuss it further.
• Figures 6, 7 are not very informative.
• Evolution of water demand satisfaction for 2056 to 2085 (lines 446-459): what are the critical periods of the year for the satisfaction of the water demand? How does the frequency of years entailing restrictions change over time?
• Figure 10: as there are thousands of small dams in the basin, please clarify how they were considered in the model.
• Discussion: please see my previous comment about the limitations of approaches based exclusively on climate projections.