The Influence of Clay Content on Cave-ins in Tank Model Tests and Monitoring Indicators of Sinkhole Formation
Round 1
Reviewer 1 Report
The paper summarizes a series of experiments of sinkhole development under different water level conditions in two different materials. The basic conclusions are: (1) groundwater levels play an important role in the speed of upward migration of the cavities after a drainage valve was opened, (2) as little as 5% clay can delay the speed of the upward expansion of the cavity, and (3) pore water pressures are immediate indicators of the soil and water drainage below and thus continuous measurements of pore water pressure can be used for sinkhole early warning.
In general, the paper is well explained and the results are convincing.
My major concern is on the presentation of the results, i.e., the figures and their captions. Figure 1 is very difficult to follow without proper captions that explain each sub-figure. In a similar manner, Figs. 3 and 4 also need specific captions for each panel, and in Fig. 4, it is also not clear what each square symbol means.
Specific comments:
Line 41: There are several important papers that deal with monitoring and early warning of sinkholes and with the effect of material properties on the time lag between cavity opening and surface collapse that should be mentioned, e.g., Nof et al (2013, 2019), Jones and Blom (2014), and Baer et al., 2018.
Line 90: change to: "…at a fixed relative density…"
Line 115: it is unclear under which water level the cavity reached the surface. Fig 1b does not help to understand that.
Lines 119-121: should appear in the methods section and not in the results.
Line 133: the delay of 60s and 30s are not evident in your figures. As far as I can see from figs 3d and 3f, the delay is about 10 seconds only.
Lines 140-141 are repeating lines 132-133
Line 144: how was pore pressure measured in the 5 cm case if the water was below the sensors?
Line 153: ….This lag may be…..
Line 154: the punching failure type should be pointed out in the figure 1(f-g) caption.
Line 176: 3c and 3e
Line 187: show in exaggerated scale that it is indeed 30-60 seconds and 100-110 seconds
Line 246: …in the assessment of the sinkhole…
Author Response
à The authors really appreciate the reviewer for the most valuable comments. This paper has been revised according to the reviewer’s comments. For the manuscript submitted, English was corrected by commercial editing service from Springer Nature. The authors believed that the minor errors will be removed in the galley proof stage. Thank you for your understanding.
à About the captions and explanation of figures, it is considered in the text. Fig. 4 is modified. Thank you.
Specific comments:
Line 41: There are several important papers that deal with monitoring and early warning of sinkholes and with the effect of material properties on the time lag between cavity opening and surface collapse that should be mentioned, e.g., Nof et al. (2013, 2019), Jones and Blom (2014), and Baer et al., 2018.
à considered in the text. Suggested papers are cited in the revised manuscript. Thank you.
Line 90: change to: "…at a fixed relative density…"
à considered in the text. Thank you.
Line 115: it is unclear under which water level the cavity reached the surface. Fig 1b does not help to understand that.
à The water level is controlled by two side chambers and fixed at a desired height (e.g., 5, 10, and 15 cm high) in small-sized tank model. However it is not so clear in the figures. This is why the authors would like to indicate the water level (blue line) in Fig. 1b. Thank you for your understanding.
Lines 119-121: should appear in the methods section and not in the results.
à considered in the text. Thank you.
Line 133: the delay of 60s and 30s are not evident in your figures. As far as I can see from figs 3d and 3f, the delay is about 10 seconds only.
à Fig. 2 shows the test results of sand. Fig. 3 shows the test results of sand and sand-clay mixtures. In Fig. 3 the delay of test start (valve opened) and sinkhole observation at the top of tank model, as shown in Figs. 1f and 1g. The authors would like to highlight the difference between observation in pore water pressure during cavity expansion and sinkhole collapse. It can be found in Figs. 3b, 3d, and 3f that there are a sudden decrease in (positive and negative) pore water pressure during the test; that is, a reduction (almost vertically) in pore water pressure is occurred in sand; however, a relatively delayed reduction is found in a sand-clay mixture due to clay content in matrix. The surface collapse is shown in Figs. 3c and 3e for approximately 30-60 seconds.
Lines 140-141 are repeating lines 132-133
à considered in the text. Removed.
Line 144: how was pore pressure measured in the 5 cm case if the water was below the sensors?
à The location of sensor is just above the water level. As indicated in the text, the sensor may have a small detection range around sensors (e.g., 5 cm). In the test, the variation in water pressure may be influenced by initial water inlet before test; however, the negative pore water pressure is generated during tests, as shown in Figs. 2 and 3 for GWL = 5 cm. The authors believed that the positive and negative (suction stress dominated) pressures are both sinkhole indicators.
Line 153: ….This lag may be…..
à considered in the text. Thank you.
Line 154: the punching failure type should be pointed out in the figure 1(f-g) caption.
à considered in the text. Thank you.
Line 176: 3c and 3e
à considered in the text. Thank you.
Line 187: show in exaggerated scale that it is indeed 30-60 seconds and 100-110 seconds
à considered in the text. Thank you.
Line 246: …in the assessment of the sinkhole…
à considered in the text. Thank you.
à Thank you for reviewing this paper and reviewer’s kind remarks. Thank you for English correction as well. The authors
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The draft entitled, The influence of clay content on cave-ins in Tank model tests and monitoring indicators of sinkhole formation, discussed the generation of sinkhole through laboratory test. The topic is interesting and practical. However, the following unclear parts must be modified.
1. P. 2, Line 66, “Jumunjin sand”, where does the sand come from? Which city in Korea?
2. In Fig. 1, add the size of the model in the figure. In addition, P. 2, Line 76, “The diamensions of the tank are 300 …included permeable porous stone.”,where is the porous stone put in Figure 1?
3. P. 2, Line 82, “had a 50 cm head difference”, what does the difference mean? Is it the difference between the two water tank or the water tank the the ground chamber?
4. P. 3, Line 93, “A high quality and easy-to-use camera…”, clarify the manufacture and the type of the camera mentioned here.
5. In Figure 1, clarify the sensors in Figure 1a, are they 5TE or pore water pressure? Where will the pore water pressure be put?
6. P. 3, Line 110, “tested groundwater levels in the ground model were selected to be 5 cm, 10 cm, and 15 cm”, it is not clear why the three heights were selected.
7. Figure 4, it is not clear the physical meaning of each symbol.
8. Find a native speaker to sharp the English of the draft.
9. The study is very interesting and the authors did a good job in the laboratory tests. However, the authors should explain the meaning of each figure in detail. So that the reader can easily understand the contribution of the study.
10. In this study, two soils (100% Jummunjin sand; 95% sand and 5% kaolinite) were studied. However, the authors should explain why only the 5% kaolinite was added and why there will be no laboratory test with maybe 30% of kaolinite?
11. If the pore pressure is a key indicator for the sinkhole, the authors should suggest how to measure the pore pressure in situ in the future.
12. With the success of the laboratory test in this study, the readers may have the following question that is there possibility to involve the following numerical methods to simulate the generation of sinkhole in the future? Discussions should be given and the references should be cited.
(1) Grain based modelling of rocks using the combined finite-discrete element method, Computers and Geotechnics, Vol. 103, pp 73-81.
(2) Simulating the failure process of the Xinmo landslide using discontinuous deformation analysis,” Engineering Geology, Vol. 239, pp. 269-281.
(3) Post-failure simulations of a large slope failure using 3DEC: The Hsien-du-shan slope, Engineering Geology, Vol. 242, pp. 92-107.
Author Response
Comments and Suggestions for Authors
The draft entitled, “The influence of clay content on cave-ins in Tank model tests and monitoring indicators of sinkhole formation”, discussed the generation of sinkhole through laboratory test. The topic is interesting and practical. However, the following unclear parts must be modified.
1. P. 2, Line 66, “Jumunjin sand”, where does the sand come from? Which city in Korea?
à considered in the text. Jumunjin standard sand is typical sand coming from Gangneung, which is located at the northeastern portion of South Korea. It is commonly used in sand-related laboratory tests, such as Toyora sand in Japan. The geotechnical properties of tested soils are as follows: specific gravity (Gs) = 2.65, D50 = 0.6 mm, Uniformity coefficient Cu= 1.5., and maximum and minimum of void ratio = 0.84 and 0.61, respectively.
References (but not cited all in the text).
Park, L.K. Suneel, M. and Im, J.C. 2008. Shear strength of Jumunjin sand according to relative density, Marine Georesrouces and Geotechnology, 26, 101-110.
J.K. Yoo and D. Park 2015. Shear strength estimation of clean sands via shear wave velocity. Journal of the Korea Geotechnical Society, 31(9), 17-27. (Korean)
Han, Y.C., Lim, H.S., and S.S. Jeong 2014. The strength and deformation characteristics of Jumunjin sand under low confining stresses. Journal of the Korea Geotechnical Society, 30(2), 33-42. (Korean)
2. In Fig. 1, add the size of the model in the figure. In addition, P. 2, Line 76, “The diamensions of the tank are 300 …included permeable porous stone.”, where is the porous stone put in Figure 1?
à considered in the text. Actually it is a porous ceramic filter plate (diameter < 0.075 mm). For this reason, the terminology is changed from stone to filter. The permeable porous ceramics are located between ground and water chambers. Thus, porous media is located in both sides of ground chamber. It is essential because the small particles of fine-grained sediments may freely move into the water chamber if there is no porous media. However it is invisible in Fig. 1. Thus porous ceramic filter plate is indicated in new Fig. 1a. Thank you.
3. P. 2, Line 82, “had a 50 cm head difference”, what does the difference mean? Is it the difference between the two water tank or the water tank the the ground chamber?
à considered in the text. Removed. The authors believed that the explanation of the water level of 5, 10 and 15 cm is enough for the readers in Applied Sciences. Thank you.
4. P. 3, Line 93, “A high quality and easy-to-use camera…”, clarify the manufacture and the type of the camera mentioned here.
à considered in the text. (Sony 4K video camera FDR-AXP35 model with 20.6 megapixels)
5. In Figure 1, clarify the sensors in Figure 1a, are they 5TE or pore water pressure? Where will the pore water pressure be put?
à considered in the text. In the caption of Figure 1. 5TE is indicated Figs. 1a and 1c. In Fig. 1a, sensors are 5TE and pore water pressure. It is explained in Fig. 1c.
6. P. 3, Line 110, “tested groundwater levels in the ground model were selected to be 5 cm, 10 cm, and 15 cm”, it is not clear why the three heights were selected.
à considered in the text. The water level is selected for 5 cm, 10 cm, and 15 cm (e.g., it is to describe a low, medium, and high groundwater level in tank) because the small-sized tank model is used. It is difficult to consider all the ground conditions, because of a relatively small-sized tank model used. However it is clear that the difference in water level may contribute to formation and collapse of sinkhole during cavity upward movement. According to the authors’ experience, the water level is important; in addition, the leakage of water (outflow quantity) is also very important. Some tests are on-going; the test results will be presented soon in the near future. Thank you.
7. Figure 4, it is not clear the physical meaning of each symbol.
à Fig. 4 is modified. A comparison of sand and clay-sand mixtures is emphasized in this paper.
8. Find a native speaker to sharp the English of the draft.
à Thank you. The authors also found some minor errors in the text in this revision process. However, it was actually edited by a commercial English editing service, by “Springer Nature”. In this revision note is not edited, but the authors believed that reviewer could understand what the authors would like to say in the text. Thank you.
9. The study is very interesting and the authors did a good job in the laboratory tests. However, the authors should explain the meaning of each figure in detail. So that the reader can easily understand the contribution of the study.
à Considered in the text. Thank you.
10. In this study, two soils (100% Jumunjin sand; 95% sand and 5% kaolinite) were studied. However, the authors should explain why only the 5% kaolinite was added and why there will be no laboratory test with maybe 30% of kaolinite?
à Considered in the text. As the reviewer mentioned, grain size (grain size distribution) dependent sinkhole formation is crucial in the assessment of sinkhole risks. In this paper, a simple method is used to compare and emphasize the existence of clay content (5% of kaolinite) in sinkhole formation. The initial ideal by authors is briefly as follows: even very small percent of clay content may help influencing and lagging the sinkhole formation. And it is true in the tank model. Based on these results, we have carried out more sinkhole tests for different clay-gravel-sand mixtures in relatively large-sized tank models (i.e., 60*30*15 and 60*30*20 cm3), which are covered with medium gravel size of 6 mm. The effect of various grain size distribution (not only for clay content) on sinkhole risks will be presented in the near future. Thank you.
11. If the pore pressure is a key indicator for the sinkhole, the authors should suggest how to measure the pore pressure in situ in the future.
à From our test results, it is clear that the variation of water pressure can be easily measured compared to other sensors, such as electrical conductivity, temperature, and volumetric water content from 5TE. However it is still very difficult to suggest the critical value of pore water pressure at the onset of cavity expansion and sinkhole formation in an in-situ condition with respect to sinkhole monitoring and early warning. We need more research works! In addition, the authors believed that this is a point value, but not an areal (regional) value. It is very applicable only for the sinkhole-prone areas resulting in severe economic and societal consequences in urban areas. The use of pore pressure in situ is possible when it is free from financial constraints. As suggested below (some landslide failure process), it may also be applicable for a relatively slow moving and reactivated mass movement, because it can be used where a disaster is expected to occur exactly. For this reason, the proposal will be left as a future study topic.
12. With the success of the laboratory test in this study, the readers may have the following question that is there possibility to involve the following numerical methods to simulate the generation of sinkhole in the future? Discussions should be given and the references should be cited.
(1) Grain based modelling of rocks using the combined finite-discrete element method, Computers and Geotechnics, Vol. 103, pp 73-81.
(2) Simulating the failure process of the Xinmo landslide using discontinuous deformation analysis,” Engineering Geology, Vol. 239, pp. 269-281.
(3) Post-failure simulations of a large slope failure using 3DEC: The Hsien-du-shan slope, Engineering Geology, Vol. 242, pp. 92-107.
à considered in the text. Suggested papers are cited in the revised manuscript.
Thank you for reviewing this paper and reviewer’s kind remarks. The authors
Author Response File: Author Response.pdf
Reviewer 3 Report
The paper deals with an interesting laboratory experiment to model the sinkhole formation mechanism in two types of soils. I think the paper is well written, clear and effective.
My only concern is on the introduction and references, which are few and do not account for the state of the art on the topic or the most recent papers.
I suggest to add the following references, (some other relevant papers can be found in their references too):
1) and 2) deal with sinkhole classifications; 3) presents some case histories also worldwide; 4) it is a numerical simulation of sinkhole triggering in urban areas.
1) Waltham T, Bell F, Culshaw M (2005) Sinkhole and subsidence: karst and cavernous rocks in engineering and construction. Springer, Berlin, p 384
2) Gutiérrez F, Parise M, De Waele J, Jourde H (2014) A review on natural and human-induced geohazards and impacts in karst. Earth Sci Rev 138:61–88. doi: 10.1016/j.earscirev.2014.08.002
3) Guarino P., Santo A., Forte G., De Falco M., Niceforo M. (2017). Analysis of a database for anthropogenic sinkhole triggering and zonation in the Naples hinterland (southern Italy). Natural Hazards, Volume 91, Supplement 1, pp 173–192. doi.org/10.1007/s11069-017-3054-5.
4) Scotto di Santolo A., Forte G., Santo A. (2018). Analysis of sinkholes triggering mechanisms in the hinterland of Naples (southern Italy). Engineering Geology, 237, 42 -52.
Author Response
1) and 2) deal with sinkhole classifications; 3) presents some case histories also worldwide; 4) it is a numerical simulation of sinkhole triggering in urban areas.
1) Waltham T, Bell F, Culshaw M (2005). Sinkhole and subsidence: karst and cavernous rocks in engineering and construction. Springer, Berlin, p. 384.
2) Gutiérrez F, Parise M, De Waele J, Jourde H (2014). A review on natural and human-induced geohazards and impacts in karst. Earth Sci Rev 138:61–88. doi: 10.1016/j.earscirev.2014.08.002
3) Guarino P., Santo A., Forte G., De Falco M., Niceforo M. (2017). Analysis of a database for anthropogenic sinkhole triggering and zonation in the Naples hinterland (southern Italy). Natural Hazards, Volume 91, Supplement 1, pp 173–192. doi.org/10.1007/s11069-017-3054-5.
4) Scotto di Santolo A., Forte G., Santo A. (2018). Analysis of sinkholes triggering mechanisms in the hinterland of Naples (southern Italy). Engineering Geology, 237, 42 -52.
à considered in the text. Thank you.
Thank you for reviewing this paper and reviewer’s kind remarks.
The authors
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The authors did a great job trying to answer almost all my questions. The draft can be accepted in the current format. Congratulations to the authors.
