the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 05 Nov 2025
Impact of surface melt and brine infiltration on fracture toughness of ice shelves
Abstract. Ice shelves are heterogeneous composites of firn, meteoric ice, refrozen melt and brine-saturated ice. The properties and distribution of these elements control ice shelf response to stress and susceptibility to fracturing. Here, we quantify how surface-melt and brine infiltration modify the Mode I fracture toughness (KIc) of meteoric ice on the Brunt Ice Shelf (BIS), Antarctica. During the 2023/24 austral summer, we recovered a 37 m core sequence from meteoric infill ice near Halley VI, where radar mapping shows continuous brine horizons at ∼ 37 m depth and line scans indicate that the upper 37 m contain ∼ 7 % refrozen melt. We combined density, salinity, temperature and grain size measurements with semi-circular three-point bending tests on samples representing (i) meteoric ice, (ii) melt-modified meteoric ice, and (iii) brine-infiltrated meteoric ice. Our results show melt-modified samples are consistently tougher than melt-free meteoric ice, with KIc increases up to ∼ 40 %. This is despite their larger grain size, indicating densification dominates over grain-size effects. In contrast, brine-saturated meteoric ice exhibits markedly lower KIc, by 14 %–34 % relative to density-matched, brine-free meteoric ice, consistent with chemical weakening and lower freezing temperatures. Our results demonstrate that as KIc varies strongly with density, salinity and depth, a spatially and temporally constant toughness value is unlikely to reproduce calving behaviour accurately. Implementing spatially and vertically variable KIc values, and understanding how ice shelf structure and composition evolves over time, is essential to improve predictions of rift propagation and calving.
- Preprint
(5613 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- CC1: 'Amendment to Affiliation and Acknowledgements', Siobhan Johnson, 06 Nov 2025
-
RC1: 'Comment on egusphere-2025-4904', Anonymous Referee #1, 02 Dec 2025
This paper presents results from fracture toughness experiments performed on ice from the Brunt Ice Shelf, which include samples from firn, refrozen meltwater, and brine-saturated layers. The experiments reveal differences in fracture toughness between the layers, which have implications for the growth of rifts and crevasses through different ice shelf materials.
Measurements of fracture toughness in ice and snow are rare, and extending the existing data record to include a wider range of ice shelf materials is an extremely valuable contribution to the literature. The results presented here will be invaluable to inform more physically realistic models of ice shelf fracture and iceberg calving. The paper is well written and clearly presented, with well chosen figures and a thoughtful discussion of the results. I have only a few minor alterations to suggest:
Figure 1: I assume the red square in panel a is the site of Halley station, but this should be mentioned in the caption. If possible, it would also be nice to indicate in panel b the path of the GPR line shown in panel c.
Lines 123-126: I don’t think you mention what the full ice thickness is at site S2, and it would be interesting to know how it compares to the 37m core length.
Section 3: On first read through it took me a minute to understand the terminology you use to label samples from different core sections (e.g. line 143, “core 35”). I’d suggest either using the terminology “core section” rather than core, or adding a sentence in section 3.1 along the lines of “Each core section was numbered sequentially on retrieval, and we use this number to identify core sections in the text (e.g. core XX)”.
Figure 4: It would help readability if you define the variables here (i.e. “T and R are the thickness and radius of the sample, a is the notch length, and S is the distance between the rollers”).
Line 181: “as quickly as possible” is a little vague – it would be useful to give an indication for how quickly this is, and why speed was important.
Figure 6: The text in some of the panels has ended up quite small and hard to read. It might help to move the colorbar, or even to separate some of the panels into a different figure. In panel a, I wasn’t clear whether there was a difference between the points colored deep blue and those colored in grey – are they all cases with no melt?
Line 224 The sentence “brine infiltration reduces KIc by 14% - 34% relative to density-matched, brine-free meteoric ice” might usefully include the qualification “at the same temperature”.
Section 5: I think it would be useful for the discussion to briefly address the effect of ice temperature and rheology on fracture propagation. Both brine infiltration and refreezing of surface meltwater are likely to increase the temperature of the ice column, and the results from Hulbe et al. (2010) suggest that areas of soft ice (such as areas of warmer ice) can reduce stress intensity at a fracture tip and reduce rift growth.
Lastly, this isn’t a suggestion to make any change to the paper, but your results made me think of Julian Scott’s paper where they found that drilling through a melt layer within the firn triggered crevasse formation – indicating that the melt layer had higher fracture toughness and had essentially been holding that area of fir together (reference below).
Hulbe CL, LeDoux C, Cruikshank K. Propagation of long fractures in the Ronne Ice Shelf, Antarctica, investigated using a numerical model of fracture propagation. Journal of Glaciology. 2010; 56(197):459-472. doi:10.3189/002214310792447743
Scott JBT, Smith AM, Bingham RG, Vaughan DG. Crevasses triggered on Pine Island Glacier, West Antarctica, by drilling through an exceptional melt layer. Annals of Glaciology. 2010; 51(55):65-70. doi:10.3189/172756410791392763
Citation: https://doi.org/10.5194/egusphere-2025-4904-RC1 -
AC1: 'Reply on RC1', Emma Pearce, 22 Feb 2026
We thank the anonymous reviewer for their careful and constructive review, and for highlighting the value of extending the fracture toughness dataset to include firn, refrozen meltwater, and brine-saturated ice from the Brunt Ice Shelf. Below we respond point-by-point to the minor suggestions and will incorporate these changes in the revised manuscript.
Figure 1: I assume the red square in panel a is the site of Halley station, but this should be mentioned in the caption. If possible, it would also be nice to indicate in panel b the path of the GPR line shown in panel c.
- We will explicitly state in the Fig. 1 caption that the red square marks Halley VI station and add the GPR transect line to panel (b) so that the location of the profile in panel (c) is visually linked to the broader map context.
Lines 123-126: I don’t think you mention what the full ice thickness is at site S2, and it would be interesting to know how it compares to the 37m core length.
- We do not know the exact thickness of the ice shelf in this location, due to the brine layer preventing GPR data from penetrating below this level, however we will add the local ice-shelf thickness at S2 in the site description as estimated from freeboard and density, which is approximately 150 m.
Section 3: On first read through it took me a minute to understand the terminology you use to label samples from different core sections (e.g. line 143, “core 35”). I’d suggest either using the terminology “core section” rather than core, or adding a sentence in section 3.1 along the lines of “Each core section was numbered sequentially on retrieval, and we use this number to identify core sections in the text (e.g. core XX)”.
- We agree that the terminology can be clarified. We will revise Section 3.1 to define our naming convention explicitly using “core section” where appropriate, and include a short sentence explaining that sections were numbered sequentially on retrieval and that this is used in the text and figures (e.g., “core section 35”).
Figure 4: It would help readability if you define the variables here (i.e. “T and R are the thickness and radius of the sample, a is the notch length, and S is the distance between the rollers”).
- We will update the Fig. 4 caption to define all geometric variables used in the equation/diagram.
Line 181: “as quickly as possible” is a little vague – it would be useful to give an indication for how quickly this is, and why speed was important.
- We will replace the phrase with a quantitative timescale.
Figure 6: The text in some of the panels has ended up quite small and hard to read. It might help to move the colorbar, or even to separate some of the panels into a different figure. In panel a, I wasn’t clear whether there was a difference between the points colored deep blue and those colored in grey – are they all cases with no melt?
- We will improve the readability of Fig. 6 by increasing font sizes and reorganising the layout to two larger graphs on top of each other. We will also correct the text, which should read ‘grey indicates 0% melt’, not blue.
Line 224 The sentence “brine infiltration reduces KIc by 14% - 34% relative to density-matched, brine-free meteoric ice” might usefully include the qualification “at the same temperature”.
- We will revise the sentence around Line 224 to include “at the same temperature”
Section 5: I think it would be useful for the discussion to briefly address the effect of ice temperature and rheology on fracture propagation. Both brine infiltration and refreezing of surface meltwater are likely to increase the temperature of the ice column, and the results from Hulbe et al. (2010) suggest that areas of soft ice (such as areas of warmer ice) can reduce stress intensity at a fracture tip and reduce rift growth.
- We will add a section in Section 5 discussing how temperature-dependent rheology may alter fracture propagation and stress intensification, and we will cite Hulbe et al. (2010).
Lastly, this isn’t a suggestion to make any change to the paper, but your results made me think of Julian Scott’s paper where they found that drilling through a melt layer within the firn triggered crevasse formation – indicating that the melt layer had higher fracture toughness and had essentially been holding that area of fir together (reference below).
Hulbe CL, LeDoux C, Cruikshank K. Propagation of long fractures in the Ronne Ice Shelf, Antarctica, investigated using a numerical model of fracture propagation. Journal of Glaciology. 2010; 56(197):459-472. doi:10.3189/002214310792447743
Scott JBT, Smith AM, Bingham RG, Vaughan DG. Crevasses triggered on Pine Island Glacier, West Antarctica, by drilling through an exceptional melt layer. Annals of Glaciology. 2010; 51(55):65-70. doi:10.3189/172756410791392763
- We agree it is relevant to the interpretation of refrozen melt layers as mechanically distinct horizons and will incorporate this paper as an additional citation in the discussion where we address melt-layer mechanical contrasts and their potential to influence rift initiation/arrest.
Citation: https://doi.org/10.5194/egusphere-2025-4904-AC1
-
AC1: 'Reply on RC1', Emma Pearce, 22 Feb 2026
-
RC2: 'Comment on egusphere-2025-4904', Lisa Craw, 09 Feb 2026
This manuscript describes a suite of experiments on samples of ice core from the Brunt Ice Shelf, examining their fracture toughness via three-point bending tests. Specifically, it examines the impact of refrozen melt and brine infiltration on the fracture toughness of natural ice. This is an interesting and worthwhile study, as previous studies have focused only on cold, dense ice, or pure laboratory-made ice, and models which incorporate calving typically assume a homogeneous value for Mode I fracture toughness. The authors show that in fact heterogeneity in the ice properties, particularly caused by melt and brine infiltration, can significantly alter the fracture toughness of the ice.
Overall it's a great paper, very easy to read and with nice results which are well presented. I have a few general comments which could be worth an additional couple of lines in the discussion to take it a bit further, but I leave it to the authors to decide if it would be over-interpreting the results to do so. All of the specific comments are very minor.
General comments:
As you mention in the introduction that most models assume a certain fixed value for fracture toughness, it would be nice to revisit this in the discussion: do you have a recommendation for how your results could be incorporated into a model? If there must be a fixed value for the whole shelf, would you change it based on these results? Or if it could be varied, what observable parameter could be used to do this?
Again, in terms of generalising your results: the Brunt Ice Shelf is quite unique in its structure. Would you expect to see the same effect on e.g. Larsen C, or is the proportion of melt to brine/ geometry of the suture zones too different?
The impact of temperature on fracture toughness is interesting, and strikes me as a little unintuitive: at higher temperatures, I would have assumed the ice would be more inclined to deform in a ductile way rather than fracture.
Do we know if the ice temperatures are higher in the brine saturated/melange areas, relative to the meteoric ice blocks? Do you think it would make a difference to the results if your experimental temperatures were more constrained?Specific comments:
line 31: 'centimetre thick' sounds like they are exactly a centimetre... -> centimetre scale?
Fig 1: I assume the bright surface at ~37m depth is the top of the infiltrating brine? It's worth labelling this just to be clear.
Fig 2: 'The spread in density from core samples is reflective of the melt within the cores': I'm not sure exactly what this means, are you referring to the lab-measured densities specifically? Or just generally the melt layers impact the density? Is the bigger spread in lab measurements because the measured sections were smaller?
Line 43: is core 35 from 35m depth? (Oh I see from figure 2 that it's ~29m. For clarity, it would help to have the depth in the text, as the core numbers are a little confusing. Same on lines 50, 59, etc)
Line 47: 'isotropic crystal fabrics', is this based on eyeballing the thin section under cross-polars? That's a fair enough way to do it, just worth clarifying.
line 71: how was surface area of melt measured? Did you take a thin section of the fracture surface, or was it visually obvious?
line 73: is this a custom-made test system, or something off-the-shelf? What is the accuracy of the load and displacement measurements?
Figure 4: I'm a little unclear on the purpose of the rollers at the sides, since it looks like the ice is sitting on the grey bits on the inside? Or is the full weight of the ice on the rollers?
line 187: I think this should be a colon instead of semicolon
line 218: missing space before referenceCitation: https://doi.org/10.5194/egusphere-2025-4904-RC2 -
AC2: 'Reply on RC2', Emma Pearce, 23 Feb 2026
This manuscript describes a suite of experiments on samples of ice core from the Brunt Ice Shelf, examining their fracture toughness via three-point bending tests. Specifically, it examines the impact of refrozen melt and brine infiltration on the fracture toughness of natural ice. This is an interesting and worthwhile study, as previous studies have focused only on cold, dense ice, or pure laboratory-made ice, and models which incorporate calving typically assume a homogeneous value for Mode I fracture toughness. The authors show that in fact heterogeneity in the ice properties, particularly caused by melt and brine infiltration, can significantly alter the fracture toughness of the ice.
Overall it's a great paper, very easy to read and with nice results which are well presented. I have a few general comments which could be worth an additional couple of lines in the discussion to take it a bit further, but I leave it to the authors to decide if it would be over-interpreting the results to do so. All of the specific comments are very minor.
- We thank Dr Lisa Craw for their careful review and assessment of the manuscript. We appreciate the constructive suggestions on how to broaden the discussion and the minor specific edits. Below, we respond point-by-point to the comments and will incorporate these revisions into the revised manuscript.
As you mention in the introduction that most models assume a certain fixed value for fracture toughness, it would be nice to revisit this in the discussion: do you have a recommendation for how your results could be incorporated into a model? If there must be a fixed value for the whole shelf, would you change it based on these results? Or if it could be varied, what observable parameter could be used to do this?
- We will add a short modelling-focused paragraph in the Discussion, suggesting the use of a GPR acquired brine map and the use of two KIC values for an ice shelf. Therefore parameterising KIc as a function of observable properties such as a brine infiltrated layer (acquired from Ground Penetrating Radar data), or known melt conditions.
Again, in terms of generalising your results: the Brunt Ice Shelf is quite unique in its structure. Would you expect to see the same effect on e.g. Larsen C, or is the proportion of melt to brine/ geometry of the suture zones too different?
- Although the Brunt Ice Shelf has an unusual structural setting, the material response we see should apply wherever these processes occur. The shelf-wide impact will, however, depend on how extensively melt layers and brine horizons are developed and how they are distributed across an ice shelf.
The impact of temperature on fracture toughness is interesting, and strikes me as a little unintuitive: at higher temperatures, I would have assumed the ice would be more inclined to deform in a ductile way rather than fracture.
- These results are focused specifically on brittle fracture, and do not account for variations in temperature associated with ductile deformation. Additionally, the temperature dependence of KIc of ice is not clearly defined. KIc has been seen to increase, with decreasing temperature (Schulson, E.M. and Duval, P., 2009) but the scatter in the data is large and the effects is small. We will add a sentence clarifying that our results focus on brittle fracture and that these relationships are weak.
Do we know if the ice temperatures are higher in the brine saturated/melange areas, relative to the meteoric ice blocks? Do you think it would make a difference to the results if your experimental temperatures were more constrained?
- We do not have in-situ measurements of the ice temperature in the brine saturated areas vs the meteoric blocks, so would be unable to comment on this at this time.
Specific comments:
line 31: 'centimetre thick' sounds like they are exactly a centimetre... -> centimetre scale?
- We will change this to “centimetre-scale thickness” (or “centimetre-scale layers”) to avoid implying an exact thickness.
Fig 1: I assume the bright surface at ~37m depth is the top of the infiltrating brine? It's worth labelling this just to be clear.
- We will label this feature explicitly in Fig. 1
Fig 2: 'The spread in density from core samples is reflective of the melt within the cores': I'm not sure exactly what this means, are you referring to the lab-measured densities specifically? Or just generally the melt layers impact the density? Is the bigger spread in lab measurements because the measured sections were smaller?
- We will clarify in the text that the spread in lab measurements are due to smaller sample sizes, with some samples containing refrozen melt, which increases that samples density relative to the average core density.
Line 43: is core 35 from 35m depth? (Oh I see from figure 2 that it's ~29m. For clarity, it would help to have the depth in the text, as the core numbers are a little confusing. Same on lines 50, 59, etc)
- We will include core depths in the text.
Line 47: 'isotropic crystal fabrics', is this based on eyeballing the thin section under cross-polars? That's a fair enough way to do it, just worth clarifying.
- Yes this is how we established fabric, and will clarify this in the text.
line 71: how was surface area of melt measured? Did you take a thin section of the fracture surface, or was it visually obvious?
- Surface melt was calculated by taking photographs of the fractured surfaces once the sample had been broken, and measuring the amount of surface area covered by refrozen melt. We will clarify this process in the text.
line 73: is this a custom-made test system, or something off-the-shelf? What is the accuracy of the load and displacement measurements?
- This is a custom made system of which we will include details of in the manuscript along with the load and displacement accuracy.
Figure 4: I'm a little unclear on the purpose of the rollers at the sides, since it looks like the ice is sitting on the grey bits on the inside? Or is the full weight of the ice on the rollers?
- We will revise Fig. 4 to remove the grey parts on the side, as this is misleading. The full weight of the sample is held by the rollers.
line 187: I think this should be a colon instead of semicolon
- This will be corrected in the updated manuscript.
line 218: missing space before reference
- This will be corrected in the updated manuscript.Citation: https://doi.org/10.5194/egusphere-2025-4904-AC2
-
AC2: 'Reply on RC2', Emma Pearce, 23 Feb 2026
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 379 | 134 | 37 | 550 | 28 | 27 |
- HTML: 379
- PDF: 134
- XML: 37
- Total: 550
- BibTeX: 28
- EndNote: 27
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
SJ's affiliation is to be amended to include affiliation with the Yusuf Hamied Department of Chemistry, University of Cambridge.
Additionally, the acknowledgements are to be amended to include SJ's funding which is a studentship granted by the Yusuf and Farida Hamied Foundation.