Dear Reviewer 1,

thank you for your very useful suggestions. We have prepared a new version of the paper where we included all of them. The point-by-point replies to all suggestions are listed below. The Editor advised me that the we will be allowed to upload the new version of the manuscript only in a successive stage when we will have received all the reviews.

We now provide a series of new data and interpretations useful to correlate the shallow information with the underground data. First, we added a new chapter of description of the internal structure of the landslide, so the chapter “2 Site description” is now divided into two subsections: “2.1 Quaternary geology and geomorphology” and “2.2 Substrate characterization”. The data of these subsections come from geological-structural field survey, logs drilled across the landslide deposits, a series of piezometers, and results of static analysis of the slope. With these data, we describe the Quaternary covers and the general architecture of the substrate of the landslide. We also describe the presence of more than one slip surface and their possible depth. This chapter is accompanied by a new figure that shows a vertical cross section through the landslide body and its substrate, completed with location of logs and potential slip surfaces.

Then we added a new chapter in the “5 Discussion”, that now is divided into two subsections: “5.1 Correlation slope deformation - lake level - rainfall” and “5.2 Behavior of the landslide and slip planes”. This latter new subsection contains a discussion on the internal behaviour of the landslide respect to the presence of different slip planes, and on the possible differential movements of the various parts of the landslide respect to an increase or a decrease of the lake level.

Following your suggestions, we also added the Latitude coordinates that were missing at the geological map, and Lat and Long coordinates at Figure 1b.

We improved the line drawing that shows correlation between sketches at different scale.

And finally, we inserted the suggestions contained in your attached pdf.

Dear Reviewer 2,

thanks for the very useful suggestions, which have been all taken into account in the new version of the manuscript. The point-by-point replies to all suggestions are listed below. The Editor advised me that the we will be allowed to upload the new version of the manuscript only in a successive stage when we will have received all the reviews.

Replies point-by-point:

GENERAL COMMENTS

The manuscript provides continuous data monitored over about three years in a site located along the eastern mountain slope of the Greater Caucasus (Georgia) overlooking the Enguri artificial water reservoir, involved in the active Khoko landslide. In particular, it reports some data about i) the landslide displacement (monitored by to two digital extensometer installed next to the head scarp), and ii) the fluctuations of the lake level.

The paper, interesting and well written, aims to provide potentially useful information for risk mitigation measures. Nevertheless, the discussion session is not able to explain the different responses monitored by the two extensometers.

Reply: we added an explanation for this in the new Discussion section “5.2 Behaviour of the landslide and slip planes”.

In particular, the Authors do not carefully argue their assumption according to which the landslide activity is almost exclusively governed by the lake levels, while the rainfall-induced direct infiltration does not significantly influence the pattern of deformation.

Reply: we now discussed the possible influence of the rainfall on the pattern of deformation in both the new sections “5.1 Correlation slope deformation - lake level - rainfall” and “5.2 Behaviour of the landslide and slip planes”.

SPECIFIC COMMENTS

Line 169. How far is extensometer n.1 from extensometer n.2 ?

Reply: 240 m, we indicated this at the beginning of the chapter “3 Methodology and instrumentation”.

Line 190. Some details regarding the about 70 mm starting value, registered on 4th November 2016, should be provided. Is it just an initial extension due to installation ? If it is so, the graph in Figure 5 should start from zero value.

Reply: Yes, it was an initial extension  due to installation; we modified Figure 5 in order to have zero as starting value.

Line 201. Such gap should be indicated in Figure 4 and the corresponding (just hypothesized) values should be reported (for instance) through a dashed line.

Reply: We modified figure 6 and 11 by changing the line segment with a dashed line.

Line 210. As already requested for extensometer n.1,  some details about the starting value of about 152 mm registered on 18 May 2017 should be provided. If it is due to installation, the graph in Figure 6 should start from zero value.

Reply: we modified Figure 6 in order to have zero as starting value, as well as in Figure 11.

Line 210. “Deformation” should be replaced (here and elsewhere in the text) by “extension”, because deformation is, of course, dimensionless.

Reply: the word has been replaced wherever necessary.

Line 240. Could you explain such different responses shown in Trench 1 and Trench 2 ?

Reply: we inserted a possible explanation for this in the new version of the paper.

Line 271 - Discussion. Such section is rather weak. In particular, it is not able to explain the different responses monitored by the two extensometers. Some properly commented figures should be added to highlight the relation between the extension rate data and the lake levels monitored during the infilling and drawdown stages. Figure 10 by itself can not  put into evidence such crucial aspect.

Reply: we added an explanation for the different responses monitored by the two extensometers in the new Discussion section “5.2 Behaviour of the landslide and slip planes”. We also put some comment/labelling in Figure 11 (previous Fig. 10) in order to better showing the correlation lake level – extension.

Lines 285-286. Such observation should be furtherly discussed. The represented daily precipitation values are not sufficient to make such observation. Rainfall accumulated over larger periods (for instance, one or more months) could agree with the observed velocity trends. Therefore, a relation between movements and direct rainfall-induced infiltration cannot be excluded.

Reply: we now made an in-depth discussion of the possible influence of rainfall on the measured pattern of extension, in both the new sections “5.1 Correlation slope deformation - lake level - rainfall” and “5.2 Behaviour of the landslide and slip planes”. We also calculated the amount of rainfall month by month to better appreciate the rain accumulated over larger periods.

Line 293. Such delay is not clear and should be discussed. In particular, I did not understand why after 29 January 2019 the rate of extension monitored at trench 1 is about 1 mm/month, while deformation monitored at trench 2 is nil.

Reply: we now discussed the possible independence of the two trenches bacause they are located on two parts of the general landslide that can move separately, also having demonstrated with the new data and geological section that there are different potential slip planes. This may also explain why trench 1 moved slowly in 2019 and trench 2 had null movement.

Figure 10. Such Figure resumes all the data shown by Figure 5, 6, 7, 8 and 9. Therefore, in my opinion, Figures from 5 to 9 could be eliminated and replaced by Figure 10.

Reply: At this stage we prefer to maintain these figures because Figure 11 (previous Fig. 10) is too dense of information, and the various lines can be better appreciated if they stand alone in each graph, especially those referring to daily rain precipitation and temperature, whose details are more difficult to be seen in Figure 11. Moreover, we are putting in Figure 11 the monthly rainfall instead of the daily rainfall, and thus the latter graph must stand alone.

TECHNICAL CORRECTIONS

Some technical corrections are reported by the attached supplement pdf file

Reply: we inserted all these corrections.

Dear Reviewer 3,

thank you for your useful suggestions. We have prepared a new version of the paper where we included all of them. The point-by-point replies to all suggestions are listed below. The Editor advised me that the we will be allowed to upload the new version of the manuscript only in a successive stage when we will have received all the reviews.

Reviewer: Data presented are interesting sine there was real monitoring. But the analysis and interpretation looks ather poor.

Reply: In the new version we added a new chapter on the inner structure of the landslide where we described more data based on logs, piezometers, geological survey, and numerical modelling. We also expanded the Discussion with the new sections “5.1 Correlation slope deformation - lake level - rainfall” and “5.2 Behaviour of the landslide and slip planes”.

Reviewer: have few major comments.  First, about the overal shape of the Khoko landslide (a in Figure 3). I'm not sure that the landslide northern boundary is correct. To my knowledge, I would drow the overal landslide more funnel-shape, with its northen boundary on the other, southern side of the "peninsula". It can be of some importance for the interpretation.

Reply: thank you for your suggested interpretation, but our Figure 3a (and the map of Figure 2) are based upon our field surveys, and also on surveys made by Soviet researchers before the lake infilling, that show the presence of scarps, sink holes and fissures in the drawn landslide body, and the drawn landslide boundaries are also based on piezometers broken by movements along the landslide slip planes, as explained in the new data section “2.2 Substrate description”. These observations allowed us to precisely draw the boundaries of this complex landslide.

Reviewer: Second, relationships between extension measured in Trench 1 and 2 and lake level shown in Figure 10 have opposit trends. While in trench 1 it followed with some delay first impoindment and, generally ignored further variations, in Trench 2 extension increased during first level drop and somehow during the second drop and following the maximal rain. Authors did not expalain it. Authers should provode more fact-based analysis not affected by some initial hypothesis.

Reply: we now made a more in-depth discussion analyzing the deformation at each extensiometer in comparison with lake level variations and rainfall values. We put in evidence the presence of different rock volumes in the landslide that can move independently, as also suggested by the presence of different slip planes (described in the new section “2.2 Substrate description”). This can explain the different behaviors recorded at the two trenches.

Reviewer: Frankly speaking, I'm not sure that there is enough data to provide any well gronded conclusions.

Reply: our paper contains four years of observations of lake level, rainfall, extension at two extensometers, and temperature, plus geological and geomorphological data coming from field surveys of the landslide area (added in the new version), plus lithostratigraphic and geotechnical data coming from more than a dozen of logs and piezometers distributed in the landslide body (added in the new version). These data are of paramount importance because this is the first monitored landslide of the Republic of Georgia, and is one nice example of monitoring of an unstable slope facing an artificial water reservoir, connected to a major hydroelectrical plant. Apart from being a useful example, the publication of these data is also necessary to arise concern about the geohazard situation at this strategic facility.

Dear Reviewer 4,

thank you for your very useful constructive suggestions and for your sketch of the landslide area that has been highly appreciated. We have prepared a new version of the paper where we included all your suggestions. The point-by-point replies to all suggestions are listed below. The Editor advised me that the we will be allowed to upload the new version of the manuscript only in a successive stage when we will have received all the reviews.

GENERAL COMMENTS 

The present work represents an interesting case study (first time for the Republic of Georgia) of a monitoring of an important landslide phenomenon facing an artificial water reservoir. Despite its uniqueness, however, the study does not seem to provide methodological or quantitative indications such as those required by such a high IF journal. Even the interpretations proposed remain generic and not adequately justified by the data collected. At present, the manuscript should be implemented and the data more thoroughly discussed and interpreted.

Reply: we have expanded the data section and the discussion, and both have been really implemented: we now present new results on the inner structure of the landslide from logs, piezometers, geological-structural mapping, and numerical modelling. We added a new figure showing a geological-structural section through the landslide body with superimposed the potential multiple slip planes resulting from numerical static slope instability modeling and log/piezometer observations. Thus, the chapter “2 Site description” is now divided into two subsections: “2.1 Quaternary geology and geomorphology” and “2.2 Substrate characterization”. We also expanded the Discussion with the subsections: “5.1 Correlation slope deformation - lake level - rainfall” and “5.2 Behavior of the landslide and slip planes”. This latter new subsection contains a discussion on the internal behaviour of the landslide respect to the presence of different slip planes, and on the possible differential movements of the various parts of the landslide also based on GPS monitoring stations located in different parts of the unstable slope, which have been now described.

SPECIFIC COMMENTS

1) Figure 3 - Observing the aerial image, the perimeter of the landslide does not seem adequately bordered. At the link below, for greater clarity, I have reported a sketch of my hypothesis based on the morphology and some characteristics of the slope, https://we.tl/t-9TlRiDhnlZ. In particular, I believe that the area is affected both by a deep phenomenon (related to the gravitational trench - DSGSD?) and by more "superficial" ones coinciding with that bordered in blue and that described by the authors.

Reply: thank you for your constructive interpretation. We have taken into account your suggestions, which, anyway, have to match with the field observations. The boundaries of the landslide area, as depicted in Figures 2 and 3a, are based upon our field surveys, and also on surveys made by Soviet researchers before the lake infilling, that show the presence of scarps, sink holes and fissures in the drawn landslide body, as well as on a series of piezometers broken by movements along the landslide slip planes and on logs. These data are now described in the new section “2.2 Substrate description”. If on one side these observations allowed us to precisely draw the boundaries of this complex landslide, on the other side we have taken into account your suggestion of possible different landslide bodies. The latter have been assessed also with the aid of GPS monitoring stations located in different parts of the unstable slope.

The latter, however, would seem to be composed of two distinct movements, with different velocity and (perhaps) type (red and orange in the sketch). This fact would also justify the different phases of activation (probably one consequent to the other).

Reply: We have taken into account this suggestion that is now included in the Discussion, and we also added a new figure where we tried to define different landslide units/bodies. In this figure we also placed the arrows representing the GPS vectors measured in different parts of the landslide.

2) Figure 10 - Based on what has been said, I believe a direct correlation between the oscillations of the lake level and the response to the estesimeters is unlikely, although these oscillations certainly represent a strongly destabilizing element. More likely a direct relationship with rainfall events; the different response time would be related to the differential movements inside the landslide body (red + orange). As is well known in geomorphology, more superficial landslides can be activated after a few days of intense rainfalls, while deeper landslides respond with delay to "seasonal" events. In this regard, the use of inclinometers inside the landslide body would have been useful.

Reply: we now made a more in-depth discussion analyzing the deformation at each extensiometer in comparison with lake level variations and rainfall values. We recalculated the rainfall values month by month (instead of daily values) to see if there is some clearer relationship with cumulated rain. We now better clarified the possible influence of rainfall on some periods of enhance extension rates, while we still think that the first strongly increase of extension rate at extensometer 1 has been influenced also by the close-in-time increase of lake level. We put in evidence the presence of different rock volumes in the landslide that can move independently, as also suggested by the presence of different slip planes (described in the new section “2.2 Substrate description”) and of GPS vectors. This can explain the different behaviors recorded at the two trenches.