The study described in this paper is substantively correct and necessary. Moreover, it surprises with the amount of work put into this study.

However, the issue is that the manuscript is now too extensive. This whole huge work is described in one huge paper, and it is something like 2 in 1, i.e. (i) results obtained for all available climate projections-CP, and (ii) for storylines. As a consequence, it is very difficult to follow, too heavy and not transparent. In addition to the volume, it is influenced by the excessive detail of some descriptions and showing in the main text elements that are not very important. For this reason, the paper is not clear and very difficult to read and understand the main issues.

1. I believe that the manuscript should be divided into two separate papers. In the paper discussed now, only the most important descriptions, results and discussions of three types of statistical analyses performed for only four storylines should be left. However, storyline No. 4 seems unlikely, so its description and results should be moved to the Supplement Material. Similarly, the RCP2.6 path is rather unlikely, so the description of the study and results for this path can also be moved from the main text to the Supplement Material.
2. A report of studies based on all the available climate projections (CP) can be successfully published in the second paper, in the discussion of which the obtained results can be related to the results for the storylines. Or vice versa - depending on the authors' decision.
3. The justification for removing descriptions of tests performed for RCP2.6 from the main text are, among others, section 3.2, Fig. 5 and Table 5. This is all unclear and too far removed from the main thrust of this paper, i.e. spatial projection of climate change on a regional scale. Considering the description in the text, it seems possible to move Table 5 to the Supplement.
4. Table 1 should rather be in the Supplement because it is additional information and does not directly concern the main research problem solved by this study.
5. line 113 – it is unclear what "strong seasonal contrast" means in story No. 3 (Purple) – this should be explained.

6. lines 140-143 - models EROS and HS1D - since these 2 models have not been taken into account in this study, why mention them (it only makes the text harder to follow).

7. Lines 144 – 160 - models EauDyssee and MARTHE - these descriptions are not necessary in the main text and can be moved to the Supplement. Only the sentence in lines 157-159 is important for the main text.

8. lines 173 -177 - excessive detail - can be moved to the Supplement.

9. Table 3 - can be moved to the Supplement because Figure 3 shows the modelled catchments well.

10. lines 193-196 - I don't understand how the SPLI values ​​were calculated. Are they based on frequency percentages? The description should be clearer, because that's the basic element of this research.

11. sections 4.1 and 4.2 - confirm that this manuscript is too burdened with too much information, analysis and discussion. All of these strays too far in different directions from the main topic given in the title. This should be an article about just one specific study - either about the storylines or about the all available CPs , depending on the authors' decision. It is also possible to propose two publications under one common main title, divided into Parts 1 and 2, respectively covering the indicated topics. 

In their paper, Jeantet et al present a study on the impacts of climate change on groundwater using the modelling plateform AquiFR. 36 combinations of Regional Climate Models and Global Climate Models and 4 storylines for 3 RCPs and different time horizons are used to feed groundwater models over approximatively half of France. The results are mainly analyzed using mapping or temporal evolutions of the Standardized Piezometric Level Index (SPLI). Particular attention is paid to 2.5, 5 or 10 years return period wet/dry events. The paper is overall well structured and organized. The high number of climate scenarios presented makes it quite difficult to present clearly and in a synthetic manner. Moreover, the high level of uncertainties associated with the final results makes the key message of the paper quite unclear. I feel like a lot of effort was put in this work and that it is worth publishing it. But I think the paper should be improved to make it more synthetic and clearer. In the following, you will find my major comments and some minors ones.

Major comments:

• The paper focuses on the impact of climate change but do not address – except at the very end of the conclusion – the strong link between water availability and water demand. One could expect agricultural and water drinking demands to change with climate change, potentially strongly impacting groundwater level. I feel like this issue should be mentioned earlier in the paper and discussed a bit more.
• The presentation of the results is hard to follow because of the 36 scenarios used and the 4 storylines. I wonder if treating the 4 storylines – that are specifically chosen to illustrate the potential contrasting futures – alone would not be sufficient to give a clear and synthetic message. Especially because the simulated groundwater levels are very uncertain with the 36 scenarios.
• I feel like treating the RCP 2.6 in the paper is not relevant. Because the current CO2 emissions trajectories make it very unlikely, and because the results for this specific RCP are so uncertain that no conclusion can be drawn. I would remove it for sake of simplicity.
• The way SPLI is computed is not clear enough for me. The way the normal CDFs are transposed in SPLI distribution should be explained.
• On all the maps, I would use a very distinct color for the part of France that were not simulated. Figure 3 shows a lot of untreated scenarios. Consequently, the maps for the treated one are very small and hard to read. I don’t know how to improve that but having larger map would be great.
• Figure 5 is really dense, and should be looked at together with Table 5 to follow the text. It’s thus quite challenging to catch all the information’s presented in part 3.2. Maybe one more argument to keep only the storylines.
• I don’t really get to what extend the results presented in Vergnes et al, 2023 are different from the MARTHE part of the submitted paper. I would appreciate if this could be discussed somewhere in the introduction.
• In part 2.2.1, it’s not very clear whether SURFEX or ISBA or both are used for the study. This part is rather confusing and unclear. I think it should be improved.
• Multi-model is advocated in the introduction. One could expect that this approach is used throughout the study. It is clearly done for the climate part but not for the groundwater modelling part. If I get it right, most of the time MARTHE and EauDyssé were not applied on the same region. And it’s hard to get where it’s been done, as all the surfaces in Table 3 are different. I think the reasons why only one groundwater model has been applied should be stated somewhere. And Table 3 should be improved so that locations where both models are applied appear clearly.

Minor comments :

• First sentence of the introduction is not clear for me. Especially “for both agricultural purposes”. Please rephrase.
• Line 83: typo – replace dot by coma
• How are the weather regimes defined?
• Line 117 : typo (Vergnes et al, …)
• Line 128 : typo - to simulate
• Line 148: “vertical exchange through aquitards” – not clear for me. Aquitards are not supposed to conduct water.
• The way unsaturated zone is treated in MARTHE should be detailed.
• Line 228: I don’t get what is meant by ”physical consistency of each storyline is conserved”
• Line 304: typo “Figure 6 the maps”. Please rephrase