the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 29 Sep 2025
Brief Communication: Hypergravity Testing of Thawing Rates in Frozen Sand
Abstract. The active layer above permafrost experiencing seasonal freeze-thaw can range from a few centimeters to tens of meters in thickness, which complicates physical modeling of this phenomenon. This study shows capabilities developed to investigate freeze-thaw in a hypergravity environment that will enable system-level experiments which tie model predictions of permafrost behavior to field observations of permafrost temperature cycling. By leveraging scaling in a hypergravity setting, this research will allow for permafrost layers to be generated on a prototype scale that capture the full thickness of the active layer on the order of tens of meters. We present preliminary results showing requirements and techniques for sample preparation, insulation, and feasible experiment run times in a 1-m radius centrifuge.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(5118 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(5118 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-2965', Anonymous Referee #1, 31 Oct 2025
-
AC1: 'Reply on RC1', Michael Gardner, 08 Nov 2025
Thank you for this comment. We fully acknowledge and agree with the reviewer's comment that this comparison should be based on sensors at similar locations to make valid comparisons. We have included an updated Figure 6 that compares sensors at the same locations. In the original comparison in Figure 6, the dense sample sensors at the top and center (RTD 21 and 20, respectively) were deeper compared to RTD 21 and 20 in the loose sample. This showed slower temperature changes in the dense sample compared to the loose sample. In the updated Figure 6, sensors at the same locations are compared (RTD 10 for the dense vs RTD 21 in the loose, and RTD 9 for the dense vs RTD 20 for the loose, location of RTD 18 is the same between tests). As shown in the updated Figure 6, this trend still holds and the discussion of soil density effects remains valid.
Similarly, Figure 4b was updated to show RTDs 10, 9, and 18 from the dense sample to compare thawing rates with those shown for the loose sample in Figure 4a. In this updated comparison, all sensors are at similar locations to make comparisons valid. The trends observed remain consistent with what was observed previously where the uninsulated samples show faster thawing from both the bottom and top of the sample, while the insulated case shows a 1-D thawing profile from the top downward. However, since the RTDs 10 and 9 are relatively closer to the surface of the dense sample compared to RTDs 20 and 21 (which were plotted in the previous Figure 4b), the timing of thawing in the updated Figure 4b occurs earlier.
As to why there were different numbers of sensors between the loose and dense samples, this is related to the number of signal conditioners that were available at the time when tests were conducted. We only had 4 signal conditioners available for the loose test, so we decided to place the RTDs in such a way that we could capture the temperature behavior near the boundaries and center of the sample. When the dense tests were conducted, we had procured additional signal conditioners so that we could get a broader spatial resolution of the time-varying temperatures within the sample. However, we did make sure to place sensors in similar locations compared to the loose tests so that we could have direct comparison between the dense and loose tests.
-
AC1: 'Reply on RC1', Michael Gardner, 08 Nov 2025
-
RC2: 'Comment on egusphere-2025-2965', Anonymous Referee #2, 30 Nov 2025
Physical testing in geotechnical and permafrost engineering holds substantial promise for advancing our understanding of complex subsurface processes. Laboratory-based physical modeling, especially under controlled conditions such as hypergravity, enables researchers and engineers to simulate and observe phenomena that are otherwise challenging to capture in the field under controlled conditions. Such testing provides invaluable data for validating analytical and numerical models, refining theoretical frameworks, and informing practical engineering solutions. The manuscript by Gardner et al. leverages these strengths and presents results that are very interesting to the readership of this journal. They really highlight the potential of hypergravity testing in cold regions engineering., However, several aspects of the manuscript could be improved to enhance clarity, rigor, and overall impact.
The study presents an approach to investigating thawing rates in frozen sand using hypergravity. The uniform conditions established in the experiments offer an ideal setting for comparison with established analytical solutions, such as the Stephen equation and the modified Berggren equation. Incorporating a more detailed discussion on a comparison between model and analytical results would significantly strengthen the manuscript and add value for readers. Additionally, further elaboration on the impact of insulation—particularly at the bottom of the samples—would provide deeper insights into how to avoid boundary effects in future physical models.
In general, the manuscript is well written and relevant references cited. The following lists additional specific comments the authors are encouraged to address:
- Abstract: The manuscript refers to active layer thicknesses of “10s of meters.” This is not realistic; while 10 – 15 m may be possible in bedrock, the development observed is typically that of a supra-permafrost talik where the frost depth no longer reaches the permafrost table.
- Line 9: When introducing permafrost, it is recommended to use the latest definition published by the CPA (Lewkowicz et al., 2025). While the current definition is not incorrect, adopting the most recent terminology is encouraged.
- Line 15: The use of the term “diseases” in reference to permafrost thawing is discouraged. Although this terminology is sometimes used by colleagues in Chinese literature, permafrost degradation should be described as a hazard, danger, or risk, rather than a disease.
- Line 38: The manuscript alternates between “field scale” and “prototype scale.” It is recommended to unify this terminology throughout the text for consistency.
- Line 41: In discussing the experimental configuration, the authors should emphasize the importance of heat fluxes rather than absolute temperature values. Ensuring consistent and uniform heat flow direction that are representative for porotype conditions is critical in physical modeling.
- Lines 45–46: The manuscript notes that freezing and thawing should occur in a one-dimensional manner to eliminate three-dimensional effects.
- Please verify and clarify the first sentence on line 61, as it appears incomplete.
- Line 81: The scaling for heat conduction is provided, but the scaling for phase change (freezing and thawing) is not addressed. Please elaborate on how phase change is reacting to hypergravity.
- Line 87 and Throughout: The term “thaw” should be used consistently throughout the manuscript, rather than “melting,” as permafrost undergoes thawing, not melting.
- Figure 3: It is recommended to add scales, annotate and label thicknesses of insulation and sample heights to improve clarity.
- Figure 5: Please provide an explanation for the observed steps in thaw penetration and why the process does not proceed gradually. Additionally, consider adding a figure showing thaw depth versus time for both insulated and un-insulated tests.
- Figure 5 and Related Discussion: The data indicate that thawing occurs from the bottom up, suggesting that the process is more complex than a purely one-dimensional problem. The manuscript should discuss the implications of this finding and the role of heat flow within the samples.
- Section 5: As previously mentioned, a comparison with analytical solutions such as the Stephen or modified Berggren equations would add significant value when discussing thaw penetration.
Citation: https://doi.org/10.5194/egusphere-2025-2965-RC2 - AC2: 'Reply on RC2', Michael Gardner, 09 Jan 2026
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-2965', Anonymous Referee #1, 31 Oct 2025
The number and elevation of sensors in dense and loose sand samples are inconsistent. For valid comparison, sensors should be placed at the same elevations and in equal numbers across both sample types. Please clarify the rationale for the current setup.
In Figure 6, direct comparisons are made between samples with different sensor positions, which can affect the validity of the results. It is recommended to justify this approach clearly, or revise the figure and analysis to focus on results from consistent sensor locations, discussing the potential impacts of any inconsistencies.
Citation: https://doi.org/10.5194/egusphere-2025-2965-RC1 -
AC1: 'Reply on RC1', Michael Gardner, 08 Nov 2025
Thank you for this comment. We fully acknowledge and agree with the reviewer's comment that this comparison should be based on sensors at similar locations to make valid comparisons. We have included an updated Figure 6 that compares sensors at the same locations. In the original comparison in Figure 6, the dense sample sensors at the top and center (RTD 21 and 20, respectively) were deeper compared to RTD 21 and 20 in the loose sample. This showed slower temperature changes in the dense sample compared to the loose sample. In the updated Figure 6, sensors at the same locations are compared (RTD 10 for the dense vs RTD 21 in the loose, and RTD 9 for the dense vs RTD 20 for the loose, location of RTD 18 is the same between tests). As shown in the updated Figure 6, this trend still holds and the discussion of soil density effects remains valid.
Similarly, Figure 4b was updated to show RTDs 10, 9, and 18 from the dense sample to compare thawing rates with those shown for the loose sample in Figure 4a. In this updated comparison, all sensors are at similar locations to make comparisons valid. The trends observed remain consistent with what was observed previously where the uninsulated samples show faster thawing from both the bottom and top of the sample, while the insulated case shows a 1-D thawing profile from the top downward. However, since the RTDs 10 and 9 are relatively closer to the surface of the dense sample compared to RTDs 20 and 21 (which were plotted in the previous Figure 4b), the timing of thawing in the updated Figure 4b occurs earlier.
As to why there were different numbers of sensors between the loose and dense samples, this is related to the number of signal conditioners that were available at the time when tests were conducted. We only had 4 signal conditioners available for the loose test, so we decided to place the RTDs in such a way that we could capture the temperature behavior near the boundaries and center of the sample. When the dense tests were conducted, we had procured additional signal conditioners so that we could get a broader spatial resolution of the time-varying temperatures within the sample. However, we did make sure to place sensors in similar locations compared to the loose tests so that we could have direct comparison between the dense and loose tests.
-
AC1: 'Reply on RC1', Michael Gardner, 08 Nov 2025
-
RC2: 'Comment on egusphere-2025-2965', Anonymous Referee #2, 30 Nov 2025
Physical testing in geotechnical and permafrost engineering holds substantial promise for advancing our understanding of complex subsurface processes. Laboratory-based physical modeling, especially under controlled conditions such as hypergravity, enables researchers and engineers to simulate and observe phenomena that are otherwise challenging to capture in the field under controlled conditions. Such testing provides invaluable data for validating analytical and numerical models, refining theoretical frameworks, and informing practical engineering solutions. The manuscript by Gardner et al. leverages these strengths and presents results that are very interesting to the readership of this journal. They really highlight the potential of hypergravity testing in cold regions engineering., However, several aspects of the manuscript could be improved to enhance clarity, rigor, and overall impact.
The study presents an approach to investigating thawing rates in frozen sand using hypergravity. The uniform conditions established in the experiments offer an ideal setting for comparison with established analytical solutions, such as the Stephen equation and the modified Berggren equation. Incorporating a more detailed discussion on a comparison between model and analytical results would significantly strengthen the manuscript and add value for readers. Additionally, further elaboration on the impact of insulation—particularly at the bottom of the samples—would provide deeper insights into how to avoid boundary effects in future physical models.
In general, the manuscript is well written and relevant references cited. The following lists additional specific comments the authors are encouraged to address:
- Abstract: The manuscript refers to active layer thicknesses of “10s of meters.” This is not realistic; while 10 – 15 m may be possible in bedrock, the development observed is typically that of a supra-permafrost talik where the frost depth no longer reaches the permafrost table.
- Line 9: When introducing permafrost, it is recommended to use the latest definition published by the CPA (Lewkowicz et al., 2025). While the current definition is not incorrect, adopting the most recent terminology is encouraged.
- Line 15: The use of the term “diseases” in reference to permafrost thawing is discouraged. Although this terminology is sometimes used by colleagues in Chinese literature, permafrost degradation should be described as a hazard, danger, or risk, rather than a disease.
- Line 38: The manuscript alternates between “field scale” and “prototype scale.” It is recommended to unify this terminology throughout the text for consistency.
- Line 41: In discussing the experimental configuration, the authors should emphasize the importance of heat fluxes rather than absolute temperature values. Ensuring consistent and uniform heat flow direction that are representative for porotype conditions is critical in physical modeling.
- Lines 45–46: The manuscript notes that freezing and thawing should occur in a one-dimensional manner to eliminate three-dimensional effects.
- Please verify and clarify the first sentence on line 61, as it appears incomplete.
- Line 81: The scaling for heat conduction is provided, but the scaling for phase change (freezing and thawing) is not addressed. Please elaborate on how phase change is reacting to hypergravity.
- Line 87 and Throughout: The term “thaw” should be used consistently throughout the manuscript, rather than “melting,” as permafrost undergoes thawing, not melting.
- Figure 3: It is recommended to add scales, annotate and label thicknesses of insulation and sample heights to improve clarity.
- Figure 5: Please provide an explanation for the observed steps in thaw penetration and why the process does not proceed gradually. Additionally, consider adding a figure showing thaw depth versus time for both insulated and un-insulated tests.
- Figure 5 and Related Discussion: The data indicate that thawing occurs from the bottom up, suggesting that the process is more complex than a purely one-dimensional problem. The manuscript should discuss the implications of this finding and the role of heat flow within the samples.
- Section 5: As previously mentioned, a comparison with analytical solutions such as the Stephen or modified Berggren equations would add significant value when discussing thaw penetration.
Citation: https://doi.org/10.5194/egusphere-2025-2965-RC2 - AC2: 'Reply on RC2', Michael Gardner, 09 Jan 2026
Peer review completion
Journal article(s) based on this preprint
Data sets
Hypergravity Testing of Thawing Rates in Frozen Sand M. Gardner et al. https://doi.org/10.17603/ds2-mwar-sp11
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 2,249 | 878 | 182 | 3,309 | 161 | 146 |
- HTML: 2,249
- PDF: 878
- XML: 182
- Total: 3,309
- BibTeX: 161
- EndNote: 146
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Michael H. Gardner
Simeon Buttery
Soo-Min Ham
Hamad Khan
Daniel W. Wilson
Jason T. DeJong
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(5118 KB) - Metadata XML
The number and elevation of sensors in dense and loose sand samples are inconsistent. For valid comparison, sensors should be placed at the same elevations and in equal numbers across both sample types. Please clarify the rationale for the current setup.
In Figure 6, direct comparisons are made between samples with different sensor positions, which can affect the validity of the results. It is recommended to justify this approach clearly, or revise the figure and analysis to focus on results from consistent sensor locations, discussing the potential impacts of any inconsistencies.