Overall, this is an interesting manuscript worth being published.

I do see, however, some room for improvement, especially in the use of the terms "resolution" and "spatial discretization". The authors appear to use different spatial resolution (10m and 50m in the horizontal direction), both of which obviously come with a different degree of geological resolution. It is unclear whether the different model results are then attributed to the spatial discretization or to the different geological resolution. To disentangle this, a grid discretization study should be conducted first, and the same grid should then be used to examine the effect of different geological resolution. I think as is, the term spatial discretization is confused with geological resolution. A poor spatial discretization may result in round-off and truncation errors, which are numerical artefacts. A fine enough grid should be free of numerical artefacts, and this grid could indeed be used to test different degrees of geological resolution. Different geological resolution may actually show different results (just like with different spatial discretizations) but the effect of numerical artefacts would be absent.

Maybe the authors could also re-think the title. It is unclear whether groundwater quantity is meant, or quality, or level, or availability? This should perhaps be clarified.

This paper examines the impact of representation of anthropogenic urban geology and spatial resolution on the simulation of shallow groundwater levels and flows. The authors developed two geological models from an existing hydrostratigraphical model by accounting for urban subsurface infrastructure and soil material and integrated them with hydrological models of 50m and 10m resolution. The effect of geologic configuration and spatial resolution are then analyzed in terms of high-water levels and particle tracking using a case study of the city of Odense in Denmark. Overall, I find the paper well written and comprehensive. It is worthy of publication.

The methods and results are well-explained, and I have no comments on them. Since this paper is submitted as part of a special issue (‘Representation of water infrastructures in large-scale hydrological and Earth system models’), I would suggest that the authors revise the introduction, discussion, and/or conclusions to bring out the broader implications of advancing representations of human-water interactions in hydrological/geological models due to increased impacts of anthropogenic interventions.

Other comments I have are minor issues to be corrected:

Line 260: Missing reference

Line 301-302: Consider revising to “and another set of parameters selected to be tied to…”

Line 302: “and” instead of “sand”

Line 315: “Eq” instead of “Ep”

Line 317: “Eq. 2” instead of “Eq. 3.2”

Eq. 2 and Line 325: Why do the weights have subscript hi, dj, and hk? Do they change according to indices i, j, k? Perhaps it will be clearer to specify the weights in Line 325 (e.g. state “w_hi = 0.45”)

Fig. 4: Refrain from using a rainbow colour scale as it could misrepresent data due to its non-linear change in hue

Fig 5: “56%” instead of “57%” in the bar plot

Line 438 – Line 441: Please check the subscripts of the model parameters. Some capitalisations are inconsistent with those in Fig. 10

Line 446-447: delete “4.3 Simulation of high-water levels.”

Line 452: “V2_50” instead of “V2_5”

Line 477: add “models” after “V2”

Line 509, 517: Check capitalization and subscript of parameter “D_r,grass” and “K_Qs2,h”

This manuscript deals with the modelling of shallow groundwater flows and levels in urbanized catchments, and highlights the impact of both the urban geology description and the spatial resolution used in the distributed model. This topic is of high interest, because the interactions between groundwater and underground constructions are important in urban soils whose features are very variable, and we need to improve our ability to simulate these complex hydrological behaviours.

The study is based on an integrated hydrological model using MIKE SHE code and this model allows a detailed representation of groundwater levels and flows. Velocity fields and then travel times may be deduced from the model; this is a real added value of this modelling application : this type of result is quite rare in the field of urban groundwater modelling and it has to be noticed. The impact of urban infrastructures in the shallow groundwater flows and level is proved through this study and this is a step forward in the urban hydrology behaviour knowledge.

The structure of the paper is basic and clear, with a first introduction section presenting the main issues related with this topic and a short state of the art dealing with urban shallow groundwater modelling, and a focus on the importance of the soil and geology description. The second section includes the case study presentation. The Geological models and the main modelling methodology adopted here is presented then and the data- modelling- and evaluation methodology adopted here. The last sections are usual, with results, discussion and conclusion.

General opinion and minor comments

This manuscript is devoted to the sensitivity of an integrated hydrological model to the urban geology, and uses 3 different representations (i.e. 3 geological models) with various consideration of the specific urban soil features. The sensitivity of the model to the spatial resolution is analysed too. For this last factor, I wonder if only two grid sizes is enough for the study of the effect of the spatial resolution.

The overall manuscript, including methods and results, is relevant and well-prepared and written. However, I have a few minor comments that could be into account in order to improve the quality of the manuscript and help the reader.

First of all, I noticed a lack of justification, especially in the Methods section. The authors did not always argue their assumptions :

- p5 l 118 : “… concrete pavement , which have an imperviousness of 75%” . How was this value estimated? Traditionnally, this kind of surface is considered as totally impervious. But I aknowledge that it may be partially pervious. But that should be explained.

- p7 l 183 : “ … and additional data on soil material in the top 5 meters”. As the modelling application is quite sensitive to the soil configuration, especially in the first meters, one can wonder where this “additional data” comes from! What kind of additional data? From drilling data? From infiltration tests?

P8 l 207-209 “ the location of roads and pipes (…) were used as proxies for the presence of excavations and trenches” What is the relevance of this assumption? Did you assess this assumption? Did you compare this proxies methodology to real data? Is it valuable only in this study case or could it be transposed in any urban catchment?

P8 214-220 – Why the SHE model was chosen here? We can understand that it is the model used by the research team, but could the authors argue why this model is appropriate to do this study? Are there any equivalent modelling tools/methods that could have been considered for this type of modelling study? Is SHE model the only one that allows to achieve the objectives of this study?

- p9 l 245. What is this surface-subsurface leakage coefficient? A parameter of the SHE model? Does it take into account the leakage in pipes, ot only the leakage from surface-subsurface? How could it be estimated?

Then, the methods section could have been improved with a graphical scheme helping the reader to understand the chosen parametrizations. This is especially needed in the 3.2.2 paragraph, because the list of the presentation of the parametrization and boundary conditions is quite long, and a scheme would be more efficient and more easy for the reader.

Finally, I have a short comment about one element of discussion : l 535-543. The sewers renovation could be a way to reduce the soil-sewer interactions and the infiltration of groundwater in sewers. As discussed by the authors, the preferential flow paths would still be present in the pipe trenches. However, I wonder if having a full renovated sewer system is not an utopy… To my opinion, there will still be some defects in the sewer system and then, as the preferential flow in the trenches remains present, the water will always find a way to penetrate in the sewers. I have the impression that this type of sewer renovation (or “non leaking pipes assumption”) is only a “modelling dream”; I am not sure it would be feasible in reality.. (especially in a economical point of view). I would appreciate that the authors re-consider this paragraph.

References

Several mistakes should be corrected :

- l57 Boukhemacha et al (2051) and Epting et al. (2008) are missing in the list of references

- l115 / l 633 : Danish Geodata Agency ?

- l197 Kristensen et al (2015) is missing in the list of references

- l 227 DHI 2017 is missing

- l 260 is specified in Fejl ! ... Like fundet / to be corrected