Dear authors,
i find the submitted paper well written and presenting interesting data. I believe the paper is a valuable contribution for The Cryosphere. In general, I agree with the authors regarding the reply to the reviewers. However, i do not fully agree that this manuscripts fully describe the potential influence of systematic errors in the resulting inversion results. I believe also that important aspects in the inversion of the data are completely overseen in the description and discussion of the results. Hence, I suggest the paper to be accepted with minor revisions and I describe below some of my concerns. Maybe the authors could try to address them, maybe within the discussion, to avoid the re-processing of data.
The paper mention the use of misfit between normal and reciprocal readings to evaluate data quality. However, such methodology can only help to identify possible erroneous measurements for data collected at a given time 'j', but the paper does not explore evolution of data quality for the monitoring. Do I understand correctly that the error parameter used in the inversion is 10% for all measurements? is this realistic? does it mean that the error did not change over time? Would not be expected to see a decrease in data quality with decreasing temperatures?
I am also struggling with the line that the data with a measured positive apparent resistivity were considered for the inversion. Did you check for possible polarity reversal due to the cross-hole measurements? We should expect some variations in the polarity, with values in the measured voltage (i.e., resistance) varying between positive and negative). Was the geometric factor modeled for all configurations to ensue that the polarity of the apparent resistivity is correctly positive?
As mentioned by one of the reviewers, the resulting inversion results show some systematically high values over certain regions. Those could be due to subsurface properties, as argued within the manuscript, but may also indicate a poor contact between the electrodes and the subsurface. Even in much easier conditions, this has been largely reported for cross-borehole ERT measurements. Here, poor contact between electrodes systematically might also result in a good agreement between normal and reciprocal readings, and typically high resistance (i.e., apparent resistivity values). I think this is a topic that is worth to be discussed in the manuscript.
I am not sure whether the running of a numerical study (as proposed by one reviewer) is completely out of the scope of the paper, as this would be an easy possibility to demonstrate that a ring electrode with a poor contact might result is large temporal variations in the measured/inverted value. However, addressing such topic as a possible reason explaining some of the features in the presented results would be important, as hinders the interpretation of possible artifacts. However, if the authors wan to skip the numerical exercise, may I recommend at least to point the readers o such inherent problem to these measurements? In case of surface electrodes, poor contact may be easily addressed by adding water, bentonite or increasing the size of the electrodes. However, in the ring electrodes in boreholes, electrodes located within voids, or having a poor contact, can not be easily assessed by the analysis of normal and reciprocal readings in a data. In one of my previous studies I proposed the definition of temporal outliers, as those quadrupoles that show a low misfit between normal-reciprocal readings at a given time, but erratic changes over time in monitoring data sets. Maybe such analysis could be performed to further evidence erroneous readings? - or maybe at least add such discussion to the manuscript for readers?
In the monitoring results, all the inversions were conducted with the same quadrupoles for all data sets? - from the manuscript, it seems that the monitoring data were independently filtered, yet inverted with the same error parameters for all readings. Is that correct? Would not be useful to run all inversions after removing all quadrupoles which are not present in all monitoring measurements? - that would help to compare inversions results with fairly the same sensitivity across the monitoring plane. However, changes in the apparent resistivity measurements over time in the same quadrupole, yet inversion of the monitoring data with the same error model may also led to overfitting in the winter measurements and underfitting in the summer measurements.

Looking forward to seeing the paper published,
Adrián Flores Orozco

Dear Authors,
the revised version of the manuscript addresses all comments from Reviewers and my list of concerns. This is a very interesting study with opens new research questions and experimental set-ups.
The manuscript is ready for publication. I am just here attaching a commented version of the paper with some minimal suggestions to the text. Maybe you can have a look into these sentences and change the formulation.
Best regards,
Adrián Flores-Orozco

Dear authors,
You have addressed all comments and suggestions from two external reviewers and mine. The three reviewers provided a positive evaluation of the manuscript and I agree that this study presents valuable data and it is an adequate contribution to TC. I recommend your manuscript for publication.
Best wishes,
Adrián Flores Orozco