The paper ("technical note") presents a practical example of using a well-known and widely used machine learning technique (neural network) for a groundwater problem. This contribution could be interesting to groundwater specialists. Referees however do no think the paper carries enough innovation, and provided a number of serious comments (and recommendations) that in their view do not allow them for recommending publishing this paper. I agree - in this form the paper is not ready for publication. However, I would like to give an opportunity to the authors to address all the referees' comments, and to rework the paper.

Dear authors
We almost there. I would still suggest to take into account the comments.
Reviewer 1 still has editorial comments.
Please note that reviewer 2 raises an important comment on "one year data" and "linear model". I suggest to answer these as well (even if he does nto insist).