Lee and co-authors present the results of geochemical analyses performed on surface/shallow blue ice samples from Elephant Moraine region, East Antarctica. These data are then used to evaluate the potentials of this region as a paleoclimate archive that extends into the mid-Pleistocene Transition (MPT). Overall, I find the manuscript interesting and worth publications after minor revisions. The merits are two-folded. First, the age (350 ka) of the shallow samples represents an exciting development. Second, the peculiar observations of the altered gas composition near the surface poses a series of questions that warrant further investigation. To better interpret the blue ice records beyond the 800 ka, it is imperative that the glaciological nature of blue ice samples be fully understood.

*Disclosure before detailed comments: I am collaborating with a few of the authors (Zheng-Tian Lu, Wei Jiang, and Guo-Min Yang) on another blue ice project, so I will refrain from evaluating the part pertaining to krypton-81 dating. Rather, my review will focus on chemical measurements such as stable water isotopes and greenhouse gases.

Line 42: “Shallow ice core drilling in BIAs has also been conducted as part of this initiative” It would be nice to give some examples.

Line 43-44: “The total area of BIAs in Antarctica is estimated to be 234,549 km2, accounting for approximately 1.67 % of the Antarctic continent (Hui et al., 2014).” This sentence seems disconnected from the text before and after it. It would be better to move this sentence to the first paragraph, after this sentence in Line 34: “… outcrops at the surface of the ice sheet in so-called blue-ice areas”

Line 52-56: The goal stated here points to the retrieval of ice core in continuous stratigraphy, but given the scope of the current manuscript it is hard to evaluate if the blue ice record at EM and RM is continuous after all. Finding another blue ice field with >1 Ma samples is a nice complement to the studies at Allan Hills. Furthermore, regardless of the stratigraphic continuity, getting the chronology is a must for any blue-ice studies, so I suggest remove the goal that you can’t reasonably accomplish with the current work.

Line 135-136: The wording here sounds as if the FID can detect CO2 and the conversion to CH4 is for the sake of sensitivity. This is not correct, as FID is highly selective to hydrocarbons (methane included). CO2, by nature, cannot be measured by FID.

Line 140: The amount of ice seems larger than what is normally required to achieve GC-FID measurements. Is this due to the lower gas content?

Line 164: "potentially contributing to differences in the provenance of blue ice between the EM and RM regions." So far there is no evidence of the provenance of blue ice in the EM and RM regions being different. I suggest move this explanation to the section where you could confidently draw the conclusion (e.g. 3.3 and 3.5).

Figure 3: I recommend re-draw this figure and divide it to three panels arranged according to cores rather than the properties being measured. The reason is that it would be useful to investigate any co-variations between CO2 and CH4, which cannot be achieved using the current version.

Line 231: The d-excess could also be used to indicate the different provenance.

Line 244-245: The question is how melting could happen a few meters below the surface. Although the maximum austral summer air temperature is −9.5 °C, direct heating during the austral summer could lead to partial melting at the surface of BIAs.

Line 248-250: The possibility of gas loss during storage and exposure would lead to more depleted Ar/N2 and O2/N2 ratios, contrary to your observations.

Section 4.1.3: Because greenhouse gas concentrations are one of the most interesting properties in polar ice cores, it is necessary to dive deeper to the issue of altered greenhouse gases. Here, modern air intrusion and melting are discussed, but other ways of producing and consuming GHGs warrant discussion. For instance, could the high CO2 be due to the in situ production? The low concentration of CH4 is somewhat more puzzling and harder to explain.

Line 268: If you revise Figure 3 and draw it according to cores, then the presence of the dust band can be marked on the new Figure. This way you could further add features of the cores to the figure to explain the altered gas composition.

Line 275: the >300 ppm CO2 in EM-B core (around 9-m depth) also warrant further investigations, especially later in the text only the top 3 meters of the data are believed to be not pristine.