the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 19 Jun 2026
Dust Record from Allan Hills Blue Ice: Towards Extending the Archive to 4000 ka
Abstract. Previous analyses of dust concentration and size distribution in ice cores are limited to the past 800,000 years; however, the ALHIC1901 ice core drilled at the Allan Hills Blue Ice Area (BIA) in East Antarctica provides a unique opportunity to examine older discontinuous records of ice ranging in age from 4000–500 ka. Here we present a discrete record of insoluble particles within ALHIC1901 from the bottom 25 m of the core. We investigated the particle mass concentration, size distribution, and mineralogy within the core to assess the preservation of dust records in BIAs with complex flow histories. We find that the insoluble particle concentrations are likely altered by entrainment of basal sediment for depths 5 m above bedrock. For shallower depths less affected by subglacial input, the record lacks expected peaks in dust concentration during glacial periods, which have been termed “long snapshots,” implying that low net accumulation rates during glacial periods at the Allan Hills BIA results in the preferential loss or attenuation of glacial ice and a corresponding bias toward the preservation of interglacial ice. The dust concentrations may also be further smoothed due to ice thinning. A subset of particles from both the upper and lower ranges of depths analyzed shows evidence of mineral weathering and/or in situ production of secondary minerals, and insoluble particle concentration correlates well with non-atmospherically derived carbon dioxide concentrations. These results highlight the importance of identifying signs of basal ice-rock interactions and/or complicated accumulation and ablation histories as these affect our interpretation of paleoclimate records preserved in ice cores from BIAs.
- Preprint
(3085 KB) - Metadata XML
-
Supplement
(3289 KB) - BibTeX
- EndNote
Status: open (until 26 Aug 2026)
- CC1: 'Comment on egusphere-2026-2434', Geunwoo Lee, 25 Jun 2026 reply
-
RC1: 'Comment on egusphere-2026-2434', Anonymous Referee #1, 30 Jul 2026
reply
Review „Dust records from Allan Hills Blue Ice: ...” by Choi et al.
Overall assessment
The paper by Choi et al. presents particulate mineral dust analyses (mass concentration, mass size distribution and some mineralogical investigations) on a new shallow ice core from the Allan Hills Blue ice area. The authors recognize that the dust record is influenced by
- temporal changes in the atmospheric composition, concentration and size distribution of mineral dust
- a potential interglacial bias of the record
- postdepositional alterations by chemical in situ weathering and neoformation of minerals in the ice
- entrainment of subglacial material into the ice
They discuss the data with respect to these issues and in how far such dust analyses could be used to separate atmospheric changes from post-depositional changes occurring in this extremely thinned and folded glaciological regime.
Overall, the results are sound, but I struggled with reading the paper, as the authors did not yet attempt a strict separation of samples affected by englacial entrainment and those sample where this can be excluded. Because of this, the discussion remains still on a descriptive level. Such a conservative triage of samples affected by only one of these factors could have made the structure of the paper more stringent and helped to separate the discussion of the issues above. At the moment, the reader has to think about the potential effects of all processes at the same time (see my more detailed comments below), which makes it difficult to distill the main take home messages.
The authors also add another section on the association of dust concentrations and elevated CO2 levels in the ice as well as the potential of carbonate minerals in contributing to elevated CO2 levels in Allan Hills Blue ice. This could be an important topic, but the short discussion on these issue does not add quantitative insight into this process yet that goes beyond what has been described in the work of Marks-Peterson. I leave it to the editor and authors to decide, whether this section should remain part of this paper or warrants its own, more quantitative paper.
Similarly, the introduction could benefit from a more stringent, linear structure to set the scene for the paper. At the moment one has the feeling that the introduction starts several times with a different aspect in the focus to motivate the paper.
In summary, I regard this paper as a relevant scientific contribution to the field, but it needs some major revisions to get the messages across more clearly.
Specific comments:
abstract and throughout the text: The authors recognize (as previous papers about the Allan Hills blue ice cores) that the record appears to be biased towards interglacials. They make low glacial accumulation rates (hence low glacial layer thicknesses) responsible for this. However, there is also glaciological evidence from deep ice cores that glacial (dust-rich) ice thins faster than clean interglacial ice (Parrenin et al., CP 2007), especially in deep, highly thinned strata. This explains at least in part the underrepresentation of glacial strata or in extreme cases could even explain a complete loss of glacial layers. Together with the formation of z-folds and finally a complete mixing of layers, this can lead to a homogeneous ice composition as found in the bottom 5 m of the presented data that is biased towards interglacials. Above 155 m still hard cuts in the dust concentration and size distribution exist, which show that there is still layering in the ice, however, it is not clearly discussed which of the samples are affected and whether these layers are mainly intrusions of deeper ice with englacial entrainment of particles (I think they are).
abstract line 32-33: Here it is stated that the dust concentrations may be smoothed by thinning. This needs some more explanation. As the particulate dust is not subject to diffusion, a more homogenous concentration record is likely only a result of the limited resolution of the data not being able to resolve the increased thinning. Note, however, that the crystal size, which is influenced by the presence of mineral dust particles sets a natural scale of variability in the impurity records.
abstract line 34: write "in situ mineral weathering in the ice matrix"
Introduction: The first paragraph motivates the paper by helping to understand the MPT, although the paper doesn't really discuss the MPT, the second paragraph motivates ice core dust studies because of iron fertilization, the third paragraph starts again by listing what else we can learn from ice cores, the fourth paragraph is somewhat redundant again. This needs some streamlining in my opinion.
line 62-63: In the sentence before, the case is made that the dust entering the ocean from above undergoes chemical alteration in the water column. This chemical alteration is crucial for its biological availability, so the pristine ice core record per se will not help to understand how bioavailable the dust really is. I would delete this sentence.
line 89: "strengthened transport of local Antarctic material and/or diminished..."
Line 95: cite also Wolff et al., CP 2022 for the use of dust as synchronization tool
Line 103: "However, with increasing age..."
Line 119-120: see my comment above on differential thinning of glacial and interglacial ice. Here also high ablation rates are mentioned. This needs some more explanation. The ice found in the Allan Hills ice core comes somewhere from upstream where there was a positive net accumulation (otherwise no signal). I can't follow the ablation argument here. Maybe, it is just a wording issue and you mean the effect on net mass balance between annual accumulation and annual ablation at the site of deposition?
line 133: It would be helpful for the outsider if it were mentioned where (roughly) the snow was initially deposited relative to the current drill site, what the accumulation and ablation conditions are at that site of deposition and how the ice flowed into the current ice core location.
line 168: "...and mass size distribution..." it is important to be exact here
line 190: here it says, you used the "midpoint of the bins". What do you mean exactly by midpoint? As these bins are equidistantly defined on a logarithmic scale, it should be the geometric mean of the bin (not the arithmetic midpoint).
line 205: " ...artificially lower measured dust concentrations and a shift of the measured size distribution to smaller particles."
line 214-217: It says that the background count for each bin (blank) was subtracted and that this correction was less than 2% of the total particle count. I assume that this blank count is different for the different bins. Please provide the number of the maximum and minimum contribution to the individual bin counts and the bin mass (a few blank counts can completely screw up the mass in the large particle bins) and tell the reader, which particles are more affected by the blank (small one or big ones).
line 223: Here it says that four selected samples were processed. Does this mean that in total SEM was performed for only 4 individual samples or did you pool 4 samples? Why were those samples selected?
line 230: Please mention the EDS spot size here (typically 1-2 microns), as this limits which particles can be measured using EDS.
Figure 3: It would be helpful to have the d18O record of the ice overlaid on this figure. It is hard to distill information from the scatter plot S1 in the supplement.
line 255: here you say 30 micron, but before you wrote the aperture was 50 micron? I also would write: "... given the aperture limit of 30 microm, but given their large size an aeolian origin of such particles can be safely ruled out (see also below)."
line 281: "produced by aeolian processes"
libe292-294: I do not understand/support the sentence: "In isolation...variability". This needs some rewording in my view.
Figure 4: This is a central Figure! First of all, it would be great for the reader to see some representative examples of the measured (unnormalized) size distributions for individual samples (not just the aggregated information in this normalized colormap). Moreover, even above 155 m one can clearly see individual samples where the typical aeolian mode around 2 microns is completely missing (suppressed by the normalization procedure). These samples are clearly not of aeolian origin (neither glacial nor interglacial, even in the Talos Dome samples that show local larger dust, one can clearly see this mode). I would recommend a stringent screening algorithm (as you suggest in your conclusions) to sort out those samples that are likely of aeolian origin, and then investigate whether their fine particle fraction changes between glacial and interglacial and, whether the mode of the small particles increases with depth as expected from in situ aggregation. The latter effect was for example seen in the bottom-most parts of the EPICA ice core by Lambert et al., 2008, 2012.
line 335-337. You performed SEM/EDS. Can you use that information to help screening the data for folds?
line 348-352: Here the Carter data are mentioned (which I couldn't access yet) but they are only shown indirectly in Figure 17. It would be helpful for the reader, if they were somehow included in the other figures 2 and 3 directly as they represent a benchmark. As that paper seems to be accepted it should be possible to include the data. I also recognized that the size distributions for ice not affected by englacial entrainment of material appear to be very similar to Talos Dome (Albani et al., CP 2012), which could be another benchmark to compare with. This is also a relatively coastal core with local and remote dust deposition but no glacial entrainment of material.
line 356: " "there is also an array"
Figure 5: Is the trend of the glacial samples (blue dots) for high mass concentrations to increasingly reduced fine particle fraction a sign of aggregation (should increase with age)?
Figure 6 and text related to it: This is an example where the analysis and discussion would benefit from conservatively screening the samples for any samples affected by folds and englacial entrainment of material.
line 403-404 and line 412-414: see my comment above, a precise wording would likely improve clarity. As mentioned above the ice must have been deposited in a net accumulation area upstream. So the ablation at the Allan Hills drill site that leads to the formation of blue ice is not exactly relevant.
line 417-420: Also here a more precise wording would help. We still have a net accumulation, otherwise we would not have a signal, but I of course agree that the net accumulation is likely to have been much smaller in the glacials than in the interglacials at the deposition site.
line 423-424: I don't understand the point of the small "subsections", please elaborate.
line 448: "expected to dissolve by cloud reworking during the long-range..."
Figure 7: add scale to each subpanel
line 475: what is meant here by "kinetic evaporation"?
line 509-522: The pH values without proper titration or correction for the dissolution of CO2 from the lab air are not sufficient. In fact, the pH of 5.6 is what we expect from lab air exposure. I would delete this paragraph.
line 527-528: "due to particle migration..."
Section 3.6: While this is an interesting topic, the discussion remains circumstantial. Importantly, the discussion in the first paragraph suggests that carbonates may be responsible for in situ CO2, however, the data by Marks-Peterson (paragraph 2) point to carbon that is isotopically lighter than the atmospheric value (respired organic carbon). If this section remains in the paper, this needs to be worded more clearly.
line 592-595: here the authors suggest a multiproxy filtering approach to distinguish samples affected by the various processes. The paper would gain of significance, if exactly this would be attempted in this paper.
Citation: https://doi.org/10.5194/egusphere-2026-2434-RC1
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 56 | 20 | 12 | 88 | 13 | 10 | 6 |
- HTML: 56
- PDF: 20
- XML: 12
- Total: 88
- Supplement: 13
- BibTeX: 10
- EndNote: 6
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Dear Authors,
Thank you for sharing this very interesting preprint. I enjoyed reading your work and found the dust analysis in blue ice and its implications insightful. I would like to offer a few constructive comments and questions for your consideration.
Regarding the discussion on size distribution changes linked to basal contamination, it might be worth considering the englacial formation of large particles. For instance, visible solid inclusions found in the EDC basal clean ice are known to form englacially and are distinct from primary basal debris (Tison et al. 2015, DOI: https://doi.org/10.5194/tc-9-1633-2015). In addition, dust aggregation might be able to affect the size distribution as discussed in the publication by Lambert et al. (2008, DOI: https://doi.org/10.1038/nature06763). While chemical alteration (line 481) and dust aggregation (line 528) are discussed later in the manuscript, explicitly addressing the potential contribution of such englacial particle growth when linking size distribution shifts to basal contamination could further strengthen this section.
It seems worth considering how the multiple parameters should be interpreted collectively. The long-term snapshots, which bias toward interglacial periods, show clearly low dust mass concentrations in the shallower sections (above 155 m depth, Figure 3) without significant basal contamination. However, these same samples still preserve distinctive variations in both dust size distribution (Figure 4, line 290) and δ18O water isotope (Supplementary Figure 1). If these layers had been significantly altered by either glacial low-accumulation or physical mixing, one might physically expect these distinct variations in size distribution and water isotopes to have been smoothed out alongside the mass concentration. Could you perhaps elaborate further on the mechanism that allows the distinct isotope and size distribution signals to survive while maintaining a consistently low mass concentration?
The SEM images are a valuable addition to the manuscript. However, adding a clearly visible scale bar to each image would help readers more accurately assess particle size and morphology.
The authors may already be aware of this, but a very recent publication on microstructural analysis by Stoll et al. (2026, DOI: https://doi.org/10.1029/2026JB033890) provides relevant insights into ice deformation and might be a helpful resource to enrich your discussion on these deeper or disturbed sections.
Thank you again for this stimulating paper, and I look forward to your response.
Best regards,
Geunwoo Lee