the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 04 Aug 2026
Updating Thermal Head of Recoverable Autonomous Sonde (RECAS) for Better Drilling Performance in Qilin Subglacial Lake Exploration
Abstract. There are numerous subglacial lakes in Antarctica, typically covered by thick ice sheets. Qilin subglacial lake is one of the largest subglacial lakes in Antarctica, lying beneath approximately 3600 m of ice. The exploration of this unique environment is of considerable scientific importance. Recoverable Autonomous Sonde (RECAS) has been proposed for exploring Qilin subglacial lake. However, drilling into the lakes with such probe may be impeded by volcanic ash or small rock clasts within the ice sheet. In addition, RECAS is incapable of withstanding the extreme hydrostatic pressure existing in the Qilin subglacial lake. To address these challenges, the thermal head of RECAS was redesigned and fabricated. Compared with the original RECAS thermal head, the updated thermal head has extra water circulation unit. Laboratory test validated that the redesigned RECAS thermal head can operate at water pressures of up to 40 MPa. Moreover, its rate of penetration (ROP) improved by 21.1 % in clean ice, by 41.1 % in dust-laden ice, and by 253.4 % in ice with debris-rich ice. The updated RECAS thermal head is expected to be utilized to drill Qilin Subglacial Lake in the coming Antarctic work season.
- Preprint
(2559 KB) - Metadata XML
-
Supplement
(86 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-4471', Anonymous Referee #1, 25 Aug 2026
-
AC1: 'Reply on RC1', Yazhou Li, 08 Sep 2026
The manuscript: Updating Thermal Head of Recoverable Autonomous Sonde (RECAS) for Better Drilling Performance in Qilin Subglacial Lake Exploration submitted by Li et al., reports on updates to an existing drilling instrument. The manuscript is well-written and generally easy to follow, also for a non-expert like me in this niche area. The English language is generally good, with only a few odd phrasings (that are still easy to understand).
The introduction is good and gives a nice overview of the field with adequate references for prior research. Obviously, this is an instrument paper with only a laboratory scale demonstration of the system, but the results presented are convincing. I have only some minor comments before I can recommend the manuscript is accepted.
Thanks for your fruitful comments. The manuscript has been carefully revised according to your suggestion.Line 110: The manuscript would benefit from an improved description of the filters used to remove dust, i.e., what materials are used. Further is also not clear to me if the dust can risk clogging the filters or does the design make dust move into the internal cavity? The text states that dust is collected in this internal cavity. How big is this cavity and is there a risk of overfilling it during deep drilling? Please clarify this in the manuscript.
The filter is made of 400‑mesh stainless steel wire screen, as described in Section 3.4.3 (near line 339).
In the design of the upgraded RCEAS thermal head, the dust was intended to be stored in the internal cavity, defined as the unused space within the outer tube that is not occupied by other components. The outer tube is a cylinder with an inner diameter of 160 mm and a length of 570 mm. Assuming that half of its internal volume is occupied by other components, the remaining cavity volume would theoretically be approximately 5727 cm³. In polar ice sheets, the dust layer thickness is typically less than 20 mm. Given a borehole diameter of 220 mm, the volume of dust encountered would be about 757 cm³. Therefore, filling the cavity to capacity would require the accumulation of seven ash layers—a condition that is rarely observed in polar ice sheets. Therefore, in our opinion, the overfilling of the internal cavity is almost impossible.
Nevertheless, a real risk of filter clogging exists. To reduce the likelihood of blockage, the contact area between the filter and water was enlarged through a double‑layer design. As illustrated in Fig. S1, the two‑layer stainless‑steel wire screen provides approximately twice the contact area of a single layer. A total of five filters are employed.Figure S1 The structure of the filter
In the revised manuscript, the following sentences have been added to the section 3.4.3 to clarify the issue.
“Nevertheless, a real risk of filter clogging does exist. To mitigate this, the contact area between the filter and the water was increased through a double‑layer design. The two‑layer screen provides approximately twice the contact area of a single layer.”
“In the future, other methods to prevent filter clogging, such as water flushing, should be explored”.
“In the design of the upgraded RCEAS thermal head, dust is intended to accumulate in the internal cavity, defined as the void space within the outer tube after the installation of other components. The outer tube is cylindrical, with an inner diameter of 160 mm and a length of 570 mm. Assuming that half of this internal volume is occupied by other hardware, the theoretical cavity volume is approximately 5727 cm³. In polar ice sheets, dust-laden layers are typically less than 20 mm thick. With a borehole diameter of 220 mm, the corresponding dust volume would be less than 757 cm³. Consequently, overfilling the cavity would necessitate the accumulation of eight such layers, which is uncommon in polar ice sheets.”Line 113 (and also 409):“When drilling in dust-laden ice, the pump must be started”. How will the operator of the system in field conditions in Antarctica know that a layer of dust has been reached? Please clarify in the manuscript.
The RECAS melt probe is equipped with built-in sensors that can measure the penetration rate in real time and transmit the data to the surface. When the operator observes a notable decrease in the rate, this indicates the likely presence of dust‑loden or debris‑rich ice, and the operator may then command the probe to activate the pump.
The following sentences are added to the revised manuscript.
“In field applications, the RECAS probe is equipped with built-in sensors that measure the penetration rate in real time and transmit the data to the surface. When the operator observes a significant decrease in the penetration rate, it can be inferred that dust-laden or debris-rich ice is present. In this case, the operator can send a command to the RECAS melt probe to start the pump”.Equation 7: Shouldn’t the x have a double dot $\ddot{x}$ above to indicate second time derivative or a d2/dt2?
In the revised manuscript, the equation has been rewritten as follows to indicate the second time derivative of x.
md2xdt2=Fg+Fz (7)Equation 9: (similar to previous comment) Using a dot above z to indicate a unit vector is a little unusual notation as this is standard physics notation for a time derivative. I suggest using a hat $\hat{z} instead.
We acknowledge that it was an error to denote the unit vector in the zdirection as z, since this notation does not represent the time derivative of z. In the revised manuscript, we have replaced it with z. Additionally, the symbols I and S have now been clearly defined.Line 152 and 187: I believe it is only necessary to mention“COMSOL Multiphysics 6.1”once.
In the revised manuscript, the phrase “COMSOL Multiphysics 6.1” has been deleted from line 152.Page 16. There are several expressions on this page, e.g. equations 17 and 19 for lambda, but it is not obvious where these expressions originate. I suggest adding a reference or two.
The method for calculating the hydraulic friction coefficient λ is widely used in engineering and is readily available in textbooks. Accordingly, we have added the following fluid mechanics reference to the revised manuscript.
Reference
Zhang, Y. (Eds.): Fluid mechanics, (second edition), Higher Education Press, Beijing, China, 273-275 pp., ISBN 9787040072723, 1999. (Text in Chinese)Line 305: There’s only a very short sentence on the supply voltage here, but I did not see a description of any electronic subsystems. I would assume that the unit contains a number of these. If so, it would be appropriate to have a description of these (or reference to previous papers with a description of these).
The updated RECAS thermal head employs a heating control system similar to the previous one, as described in Yu et al. (2021). The voltage of the lower thermal head is regulated by a self‑developed 800 VAC solid‑state power regulator. It uses the cycle power regulation method, which controls the conduction of 50 sinusoidal half‑cycles per second (at 50 Hz) with a 2% resolution, thereby adjusting the effective power applied to the cartridge heaters in the thermal head. Since the thermal head has good heat dissipation and high temperature tolerance, its power control is simply open‑loop or proportional regulation, without the complex fuzzy PID closed‑loop algorithm. During downward drilling, the lower thermal head can be set to full power or reduced power as required to satisfy different penetration rate needs.
In the revised manuscript, the following sentences have been added.
“Similar to the previous design, the updated RECAS thermal head uses a self‑developed 800 VAC solid‑state regulator. This regulator employs cycle power regulation, controlling the conduction of 50 sinusoidal half‑cycles per second to adjust the effective power, thus enabling full‑power or reduced‑power operation to match different penetration rate requirements (Yu et al., 2021)”
Reference
Yu, H., Zhu, T., Jiang, X., Tang, Y., Li, X., Li, C., Huang, S., Shi, J., Sun, Y., Talalay, P., Fan, X., Li, X., Li, Y. and Peng, S.: Recoverable Autonomous Sonde for subglacial lakes exploration: heating control system design, Ann. Glaciol., 62, 280-292, https://doi.org/10.1017/aog.2021.5, 2021.
Line 359: The increase in rate of penetration is given as“approximately 21.8%. Based on the good, but somewhat“noisy”results in figures 12, 13, and 14 (which are all based on only a single experiment) the quoted increase seems to contain a bit too many digits. I believe it would be fairer to state 22 % or even round to 20 %.
In the revised manuscript, all percentages that describe the increase in penetration rate have been rounded to integers, with no digits retained after the decimal point. As examples, the values 21.8%, 33.3%, 41.1%, and 253.4% were rounded to 22%, 33%, 41%, and 253%, respectively.
-
AC1: 'Reply on RC1', Yazhou Li, 08 Sep 2026
-
RC2: 'Comment on egusphere-2026-4471', Anonymous Referee #2, 12 Sep 2026
-
AC2: 'Reply on RC2', Yazhou Li, 14 Sep 2026
This paper presents results of quantitative though very limited laboratory tests on a melt head intended for the RECAS probe, which incorporates water jetting to penetrate ice with embedded layers of sediment and small (cm-scale) rocks. (The paper also preliminary calculations of flow rates to suspend sediment particles and required jetting-pump pressure.) While jetting for passage of an electrically powered probe through debris-laden ice has long been reported (the authors’ references I would add Rado et al.,1987), quantitative data (e..g., on flow rates required for debris with specific characteristics) seem to be, at best, very rare. This paper therefore contributes to work not only with the RECAS melt head but also more generally, despite its limited results. I therefore recommend publication, but also recommend revision to address the following points, so as to make the paper contribute more generally. With revisions to address these points, I think the paper would make a worthwhile contribution to the literature. (I will not reiterate the points made by anonymous reviewer 1, with which I agree and which the authors have already addressed).
Thank you for your valuable comments. They are very helpful for us in improving the manuscript according to your suggestions.
In the revised manuscript, the reference suggested by the reviewer regarding water jetting in an electrically powered probe has been added to the introduction section. The following sentence “Since the first HWDDSC was developed in 1972, a wide range of such drilling systems have been developed and deployed globally (Gillet, 1975; Hantz and Lliboutry, 1983).” has been changed to “Since the first HWDDSC was developed in 1980s, a wide range of such drilling systems have been developed and deployed globally (Rado et al., 1987)”.
Added reference
Rado, C., Girard, C., and Perrin, J.: Electrochaude: A self-flushing hot-water drilling apparatus for glaciers with debris, J. Glaciol., 33, 236–238, https://doi.org/10.3189/S0022143000008741, 1987.
Deleted references
Gillet, F.: Steam, hot-water and electrical thermal drills for temperate glaciers, J. Glaciol., 14, 171–179, https://doi.org/10.3189/S0022143000013484, 1975.
Hantz, D. and Lliboutry, L.: Waterways, ice permeability at depth, and water pressures at Glacier d'Argentière, French Alps, J. Glaciol., 29, 227–239, https://doi.org/10.3189/S0022143000008285, 1983.(1) The authors base their expectations of the volumetric concentrations and characteristics of debris in ice above their target lake on observations of ash layers in cores well above the bases of ice sheets, and on (the few) inclusions observed in accreted ice above or bordering Lake Vostok. However, at ~3600 m depth, the ice sheet upstream of their intended target lake is certainly wet at its base. Ice flow over a wet bed entrains much denser concentrations of sediment and rock than those in the authors’ tests, as shown for example in data from Byrd Station (Gow et al.,1979) and in the WISSARD and SALSA project experiences. The authors, as well as others interested in melting probe entry into subglacial lakes, should therefore work toward tests with much denser debris concentrations. The first essential step to evaluate the current results and plan for further tests is to report in Section 4.3 not only the dust-laden ice layer thickness, but also the material (silica?), volumetric concentration or mass fraction of dust in the layer (especially), and dust particle size distribution (not just minimum particle size)
Thanks for your valuable and insightful comment. We agreed that denser concentrations of sediment and rock may exist at the base of ice sheet, as documented at Byrd Station by Gow et al. (1979), where basal ice contains debris weight percentages of approximately 12–16% and clasts up to ~8 cm.
In this study, the dust-laden ice sample was prepared using river sand, which consists primarily of silicon dioxide (SiO₂). The dust in the ice layer had a mass fraction of approximately 21% and a particle size ranging from 250 to 425 μm. To clarify the information, the original sentence was rewritten as “In preparation, dust particles with size ranging from 250 μm to 425 μm were selected to ensure conservative and reliable test conditions, as this size exceeds the typical particle sizes found in Antarctic ice sheet. The dust‑laden ice layer was prepared using river sand, which consists primarily of silicon dioxide (SiO₂). With a thickness of approximately 2 cm and mass fraction of approximately 21%, the dust-laden ice layer was located about 30 cm below the ice surface.”.
In general, the artificial dust-laden ice layer has a relatively high dust concentration and a larger dust particle size than those found in ice cores. However, such dust-laden ice layers may occur successively at the base of the ice sheet, leading to cumulative hindering effects on ice penetration. In our study, the multiple dust-laden ice layers were not prepared, which is a clear limitation of the present study. In the revised manuscript, we have added a discussion of this limitation and clarified that our results should be extrapolated with caution to field conditions with multiple dust-laden ice layers. The following are the added sentences, “Notably, multiple dust-laden ice layers may exist within the ice sheet, and their potential hindering effects on ice penetration may accumulate. In this study, the updated RECAS thermal head was tested only in a single dust-laden ice layer; therefore, its drilling performance in multiple dust-laden ice layers should be carefully assessed.”
In addition, an outlook section has been added at the end of the manuscript to remind relevant researchers that further in-depth research is needed in the future. The following is the added outlook section, “However, the updated RECAS thermal head has never been deployed in a real ice sheet overlying a subglacial lake. For future work, tests with multiple dust-laden ice layers and significantly higher debris concentrations, with larger volumetric fractions and larger rock sizes, are recommended to better approximate the extreme basal ice conditions above subglacial lakes”.(2) In the case of rock inclusions (Section 4.4), it must be expected that ice above the intended lake will have debris layers with many rocks in close proximity (many more than 5). To begin to evaluate the significance of the results with so few rocks, the first question is whether melt head penetration at the specified jetting rate, and the mechanism the authors suggest (melting of a cavity large enough for the rocks to move up the sidewall of the probe as it descends), are in fact repeatable, or whether other results are sometimes obtained – the authors presently seem to have presented results so far only from a single trial. If the current results are repeatable, it would then be essential to begin to explore the relationship between successful penetration and (1) jetting rate; (2) mass fraction of rock embedded in an ice layer; and (3) rock size. I would urge the authors to add data along these lines to the paper.
We are confident that the results and mechanism are repeatable and reliable. In fact, we conducted two laboratory tests with rock clasts. The first is presented in the manuscript; the second used transparent ice blocks to directly observe water cavity formation (Figure S1). The transparent ice block had dimensions of 1 m × 0.5 m × 0.5 m.Figure S1 Test of updated RECAS thermal head with transparent ice block
During the test, the following procedure was used. First, the pump of the updated RECAS thermal head was started to drill clean ice and observe the formation of a water cavity. Second, the pump was turned off, and drilling continued by thermal melting until the cavity disappeared. Third, the thermal head was lifted, and 10 pebbles with diameters of 0.5–1.0 cm were added to the bottom of the borehole. Finally, the thermal head was redeployed to the borehole bottom, and the pump was restarted to observe how the water cavity formed. Same as the first test, during the second test, the weight on the thermal head was stabilized at 150 N, the heating power was set to 7 kW, and the ice sample temperature was maintained at −15 °C. Additionally, the flow rate of the magnetic gear pump remained at 2.5 L min-1.
As shown in Figure S2(a), a conical water cavity formed during drilling in clear ice, with both its depth and maximum diameter approximately 60 mm. After the pebbles were added, the cavity still existed, and all the pebbles settled at its bottom (Figure S2(b)). In this case, the cavity had almost the same dimensions as the one without pebbles.Figure S2. Photograph of the water cavity: (a) without pebbles; (b) with pebbles. (In the revised manuscript, Figure S2 has been added as Figure 15.)
To clarify this, the following sentences have been added to the revised manuscript. “To clearly observe the formation of the water cavity, a transparent ice block with dimensions of 1 m × 0.5 m × 0.5 m was used for supplementary drilling. In the secondary test, the pump of the updated RECAS thermal head was first activated to drill through clean ice and observe water cavity formation. The pump was then switched off, and drilling continued by thermal melting until the cavity disappeared. The thermal head was subsequently lifted, and 10 pebbles (0.5–1.0 cm in diameter) were placed at the bottom of the borehole. Finally, the updated RECAS thermal head was lowered back to the borehole bottom, and the pump was restarted to observe the reformation of the water cavity. In the second test, the weight on the thermal head was stabilized at 150 N, the heating power was set to 7 kW, and the ice sample temperature was maintained at −15 °C. Additionally, the flow rate of the magnetic gear pump remained at 2.5 L min-1”
“As shown in Figure 15(a), a conical water cavity formed during drilling in clear ice, with both its depth and maximum diameter approximately 60 mm. After the pebbles were added, the cavity still existed, and all the pebbles settled at its bottom (Figure 15(b)). In this case, the cavity had almost the same dimensions as the one without pebbles. Clearly, the formation of a water cavity is the main reason for eliminating the influence of pebbles on ice drilling”.
We agreed that different rock mass fractions and rock sizes require different flow rates to generate water cavities of different sizes. In addition, the accumulation of rock during drilling may also necessitate a larger water cavity. Theoretically, the larger the water cavity, the more favorable it is for drilling through debris-rich ice. However, a large water cavity also requires a large downhole pump, heated meltwater, and specialized spray nozzles, which is beyond the scope of this manuscript. In the future, dedicated research may be conducted on this issue. It should also be noted that the updated RECAS thermal head was designed to penetrate debris-rich ice with low rock mass fractions and small rock sizes. If the rock size reaches 8 cm or even larger, and the debris-rich ice layer is 4.83 m thick, as at Byrd Station (Gow et al., 1979), it would be difficult for the RECAS thermal head to drill through such an ice layer, even with a high flow rate. In this case, a conventional hot-water drill may be the only clean way to access a subglacial lake.
In the revised manuscript, the following sentences have been added to explain this issue. “Notably, the updated RECAS thermal head was designed only for debris-rich ice with low pebble mass fractions and small pebble sizes. It would be difficult for the thermal head to drill through thick debris-rich ice with large rock clasts, as at Byrd Station (Gow et al., 1979), where the layer is 4.83 m thick and the largest clast is 8 cm. In the future, dedicated research could be conducted to optimize the flow rate and nozzle structure to create a desired water cavity that can collect as many rock clasts as possible.”(3) Ice temperature is likely also important to understanding how general the results presented here may be. I looked for but did not find any data in the paper on ice temperatures during the penetration experiments。
The ice temperatures during testing were presented as follows: “clean ice samples were fabricated in a freezing container at −20 °C” (Section 4.2); “the dust-laden ice sample was frozen at −15 °C” (Section 4.3); and in Section 4.4, “each barrel was first filled with tap water to 50 % of its volume and then frozen at −15 °C for 24 h to form a dense ice layer”.
To make it clearer, the following sentences have been added
“The ice sample was maintained at −20 °C” (Section 4.2)
“Specifically, the weight on thermal head stabilized at 150 N, and the heating power was set to 7 kW, and the ice sample temperature was maintained at −15 °C” (Section 4.3)
“Additionally, the debris-rich ice sample was maintained at a temperature of −15 °C” (Section 4.4)Finally, one minor editorial note: the various drilling and melt probe efforts summarized in Section 1 have been reviewed comprehensively, along with many others, by Talalay (2020). I suggest referring to that work.
References
Gow, A.J., S. Epstein, and W. Sheehy (1979), “On the Origin of Stratified Debris in Ice Cores from the Bottom of the Antarctic Ice Sheet”, Journal of Glaciology 23(89), 185-192, doi:10.3189/S002214000029828.
Rado, C., C. Girard, and J. Perrin (1987), “Electrochaude: A self-flusing hot-water drilling apparatus for glaciers with debris”, Journal of Glaciology 33(114), 236-238.
Talalay, P.G. (2020), Thermal Ice Drilling Technology, Springer Nature Singapore Pte Ltd, ISBN 978-981-13-8847-7 (published electronically as ISBN 978-981-13-8848-4, doi.org/10.1007/978-981-13-8848-4).
In the revised manuscript, all the three references have been added.
References
Gow, A.J., Epstein, S., and Sheehy, W.: On the origin of stratified debris in ice cores from the bottom of the Antarctic ice sheet, J. Glaciol., 23, 185–192, https://doi.org/10.3189/S002214000029828, 1979.
Rado, C., Girard, C., and Perrin, J.: Electrochaude: A self-flushing hot-water drilling apparatus for glaciers with debris, J. Glaciol., 33, 236–238, https://doi.org/10.3189/S0022143000008741, 1987.
Talalay, P.G. (Eds.): Thermal ice drilling technology, Springer, Singapore, ISBN 9789811388477, 2020.
-
AC2: 'Reply on RC2', Yazhou Li, 14 Sep 2026
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 82 | 40 | 17 | 139 | 10 | 9 | 8 |
- HTML: 82
- PDF: 40
- XML: 17
- Total: 139
- Supplement: 10
- BibTeX: 9
- EndNote: 8
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The manuscript: Updating Thermal Head of Recoverable Autonomous Sonde (RECAS) for Better Drilling Performance in Qilin Subglacial Lake Exploration submitted by Li et al., reports on updates to an existing drilling instrument. The manuscript is well-written and generally easy to follow, also for a non-expert like me in this niche area. The English language is generally good, with only a few odd phrasings (that are still easy to understand).
The introduction is good and gives a nice overview of the field with adequate references for prior research. Obviously, this is an instrument paper with only a laboratory scale demonstration of the system, but the results presented are convincing. I have only some minor comments before I can recommend the manuscript is accepted.
Line 110: The manuscript would benefit from an improved description of the filters used to remove dust, i.e., what materials are used. Further is also not clear to me if the dust can risk clogging the filters or does the design make dust move into the internal cavity? The text states that dust is collected in this internal cavity. How big is this cavity and is there a risk of overfilling it during deep drilling? Please clarify this in the manuscript.
Line 113 (and also 409): “When drilling in dust-laden ice, the pump must be started”. How will the operator of the system in field conditions in Antarctica know that a layer of dust has been reached? Please clarify in the manuscript.
Equation 7: Shouldn’t the x have a double dot $\ddot{x}$ above to indicate second time derivative or a d2/dt2?
Equation 9: (similar to previous comment) Using a dot above z to indicate a unit vector is a little unusual notation as this is standard physics notation for a time derivative. I suggest using a hat $\hat{z} instead.
Line 152 and 187: I believe it is only necessary to mention “COMSOL Multiphysics 6.1” once.
Page 16. There are several expressions on this page, e.g. equations 17 and 19 for lambda, but it is not obvious where these expressions originate. I suggest adding a reference or two.
Line 305: There’s only a very short sentence on the supply voltage here, but I did not see a description of any electronic subsystems. I would assume that the unit contains a number of these. If so, it would be appropriate to have a description of these (or reference to previous papers with a description of these).
Line 359: The increase in rate of penetration is given as “approximately 21.8 %”. Based on the good, but somewhat “noisy” results in figures 12, 13, and 14 (which are all based on only a single experiment) the quoted increase seems to contain a bit too many digits. I believe it would be fairer to state 22 % or even round to 20 %.