The paper should be carefully revised to answer the reviewers' comments. In particular, I stress that the comments by referee #2 should be seriously considered. As a matter of fact, I think that the authors' answers to some of the comments by referee #2 are not fully satisfactory.

The reviewers provided different assessments of the scientific significance and quality of the paper, ranging from Excellent to Fair, and also of the presentation quality. However, both reviewers suggest acceptance of the paper after minor revision or technical corrections, even if some important criticisms are raised.
I carefully read again the paper and I think it requires some work before reaching the quality threshold to be accepted for publication on HESS.
In particular, I expect that the Authors improve the paper, by taking into account and properly answering both the reviewers' comments and the questions that I list below.
1) The presentation and the language should be improved. I provide several suggestions in the attached marked version of the manuscript, where some of the most obscure (if not incorrect) sentences are annotated.
2) From the physical point of view the choice of a 2D model is not supported by solid arguments. In fact, a 2D model assumes that the quantities are constant along the direction perpendicular to the modelled cross section, i.e., parallel to the coast line. Is this realistic? Does the conduit network have a large extension along that direction? How are the results affected from 3D flow and transport in the porous matrix?
3) I fully agree with the remark by Referee #2 that the discussion of the results is very long and difficult to follow.
4) I fully agree also with the remark by Referee #2 about CSS and weighting.

The paper has been revised and it has been significantly improved.
The language has been strongly improved, although an editorial revision will be necessary before publication.

From the scientific point of view, a couple of points remain open.
1) How are the weights appearing in (2) computed?
2) I do not agree with the authors that "the assumption of two-dimensional model is reasonable and sufficient for understanding the parameter sensitivities, since the horizontal exchange flux between the conduit and the surrounding porous medium is relatively small". If this were true, flow and transport in the conduit would be independent from the porous matrix and the model could be further simplified.

A final technical comment: please rephrase and explain the sentence at line 819 to 820 "the mixing is mostly due... in this study."

The paper has been revised, by including some information and comments, following the comments to the last version. Unfortunately, two basic scientific concerns remain.
1) About the use of a 2D model, I do not agree with the sentence at lines 280 to 281 ("The exchange fluxes... are ignored"). In particular, I cannot find a proof of this statement in the paper and I simply suggest to erase the sentence.
2) The weight appearing in equation (2) is "set as 1.0 equally for the 11 evaluated locations (column #25 to #75 with an interval of 5 cells) for salinity and head simulations in this study" (see lines 172 to 174). In other words the weight omega_{i,i} for head simulations is 1 m^2 and 1 PSU^2, respectively for head and salinity. I recall that head varies between 0 m and 1.52 m in the simulated domain, whereas salinity varies between 0 PSU and 35 PSU. Therefore the weight does not take into account the difference in measurement units and the variability of the two quantities. In other words, with this choice (same weight for head and salinity simulations) dss is not really "dimensionless", because it does not take into account the different measurement units and therefore the different numerical values of head and salinity. This choice must be physically-based and explained, otherwise the scientific merit of the paper will be significantly lowered.

A technical remark. I recall that SI measurement units should be used throughout the whole paper, whereas other units still appear somewhere in the text and in many figures.

In conclusion, I hope that the Authors can fix the paper with a proper answer to the above mentioned comments. In particular, I wish that the Authors provide a physically-based and convincing motivation of the choice of the unitary value for the weight in equation (2). If a clear and convincing answer to point 2) will not be given, so that the problem remains unresolved, I am sorry, but my opinion is that the paper could not be published. However, in that case, before taking my final decision, for the sake of fairness, I will ask the reviewers to reconsider the paper again for this specific point.

The Authors corrected the definition of the weight for the computation of "dss", which is now clear and more accurate than in the preceding version. Also, the discussion of the 2D approach is richer, more precise and satisfactory than the previous one.
The paper has now reached a quality sufficient for publication on HESS, even if it eventually needs a careful copy-editing.