Review:

Modelling the evolution of permafrost temperatures and active layer thickness in King George Island, Antarctica, since 1950

Joana Pedro Baptista, Gonçalo Brito Guapo Teles Vieira, Hyoungseok Lee, António Manuel de Carvalho Soares Correia, Sebastian Westermann

This study investigates permafrost dynamics in an ice-free region of Barton Peninsula, located on King George Island in the Antarctic Peninsula. The authors employ the CryoGrid Community Model to simulate ground temperature at a bedrock site where a borehole was recently installed. This allowed the model to be validated using observational data from 2020 to 2022. Following this validation, the authors used adjusted ERA5 data to drive the CryoGrid Model for a long-term simulation spanning 1950 to 2022.

The paper is well-structured and meticulously prepared. The authors implement a rigorous validation approach, followed by a thorough evaluation of the results. Additionally, the discussion provides a well-balanced critical analysis of the study's findings. I strongly endorse this paper for publication.

However, there is a broader critique of the paper. In several sections, it becomes challenging for readers unfamiliar with the CryoGrid Model to follow the discussion. I recommend that the authors revise certain paragraphs where specific features of the CryoGrid community are mentioned without sufficient explanation, as these may be unclear to readers without prior knowledge of the model.

Here are a few examples where readers may struggle:

1. Line 130: The paper states that a "snow factor" of 0.3 is used, but this term is neither defined nor explained in the context of the CryoGrid Model. Clarifying its meaning would enhance reader comprehension.
2. Line 132: The text repeatedly references ground column classes, such as "GROUND_freeW_seb_snow." If these classifications are included, they should be clearly explained for those unfamiliar with CryoGrid. Alternatively, consider removing them if they are not essential.
3. Lines 179-182: The N-factor is introduced abruptly without explanation of its meaning or relevance. Table 1 also includes "Freezing N-factor" and "Thawing N-factor," but their significance is unclear. The authors should explicitly justify the inclusion of N-factors, possibly in the model description section.

Although I am not a CryoGrid specialist, I will briefly comment on the N-factors. If the paper relies on N-factors as originally introduced by Lunardini (1978), it is important to note that this approach is not ideal for modeling the TTOP method. N-factors require careful long-term calibration, as n_f and n_t exhibit considerable annual variability. In general, N-factors are highly empirical and often used alongside the R-factor to calibrate surface and TTOP temperatures. If the CryoGrid Model applies N-factors in this study, it is likely that the freezing factor n_f has not been set correctly due to its large variability. This could explain the challenges in simulating thaw depth, as N- and R-factors likely influence freeze-back in autumn (see figure 9). Overall, I recommend that the authors illustrate the impact of N-factors within the study by including figures showing how N- and R-factors vary over time.

I have some minor small comments, which are listed below.

Minor comments:

Line 20: … shoulder season. Please clarify which time period you really address.

Line 29: 70,000 km2. Please use superscript for km²

Line 33: The increase of the MAAT is 3.4°C. Please mention also the date and the original MAAT since the increase of MAAT has occurred. Please clarify.

Line 52-56: This is your validation strategy. Please make this clearer. The first hypothesis is a direct validation of your local borehole conditions, and the second hypothesis is an indirect comparison between MAAT values and your permafrost borehole observations. Therefore, the second hypothesis can only be fulfilled if the first hypothesis delivers acceptable results. Please clarify the explanation of the hypothesis.

Line 69: the MAAT mentioned here is -2.2°C for 1969 to 2022. If we calculate the warming rate of 0.09°C per year for the given time period then a warming of 4.77°C will result, which does not correspond to the value increase of MAAT of 3.4°C which is given in the introduction section? Please clarify.

Line 75-78: In the first sentence you present different values for sea level (-0.9°C) and unglaciated peaks (-3.2°C) but in the next sentence you give values of -3.6 to -0.8°C not explaining on which altitudes these values for MAGST are measured. Please clarify. Please also mention in the paper at which altitude the borehole was drilled.

Line 81-82: do you have a reference for this statement?

Line 89: This means that you have not a measured stratigraphy at the borehole site? When you did the drilling, you have not established a stratigraphy? Please clarify?

Figure 3: I would like that you integrate the expressions like validation and calibration within your model workflow to make the overall understanding clearer.

Lines 118, 127, 132: please explain these classes such as “GROUND_freeW_seb_snow” or remove it from the text because they do not explain the reader anything if you are not working with CryoGrid.

Line 130: what means a ‘snow factor of 0.3’. Please clarify.

Line 134: jhc please use subscript j_hc.

Line 179-182: Here the authors suddenly introduce the N-factor. It was nowhere before presented in the paper. Please read my content about the N-factors mentioned above.

Line 180: FS and TS are not officially introduced in the text, please define it at the first place in the text and then afterwards only use the introduced abbreviation.

Figure 5: Please show air temperature, and borehole temperature in 0.02, 0.4, 1.2, 2.0 m depth in one figure and 3 to 13 m depth in another figure with different axis to better show the variabilities.

Figure 6: Please use other scale of the y-axis to see the variabilities of the borehole temperatures within 6, 10 and 13 m better.

Line 247: only ground temperatures measurements are validated in this chapter, which is ok. However, there is no validation of any surface energy balance variables as you have not measured them. Therefore, please change the title here to Validation of ground temperatures based on the model results.

Figure 8: Please use other scale of the y-axis to see the variabilities of the temperatures within 6, 10 and 13 m better.

Line 325: Please remove the word ‘thin’ as it is always a relative statement here and just focus on the real value of 1.6 m.

Line 326-327: I do not understand this sentence: what means snowfall with a decrease from 15 m to 13 m? Please clarify.

Line 341: Maybe the word ‘intensity’ is better than the word ‘severity’.

Line 364-365: I do not fully understand this sentence. If you had more snow in spring and early summer it would actually protect permafrost thawing as at the time during the year when you have the highest shortwave incoming radiation and there is still snow on the surface, the energy is used to melt the snow and not to heat up the surface ground cover. Therefore, it is not clear what you want to express with this sentence? Please clarify.

Line 423: change: for permafrost since inhibit to distinguish

Reference:

Lunardini, V. J., Theory of n-factors and correlation of data, in Proceedings 3rd International Conference on Permafrost, Edmonton, Alberta, Canada, 1978, Volume 1, National Research Council, Canada, p. 40-46.

In this paper, the authors provide new information on permafrost temperatures and active layer thickness at a specific location within the northern part of the Antarctic Peninsula. The paper provides strong evidence of the climate and permafrost warming during the last 50 years and reports on an increase in the rate of this warming after 2025. The topic of this paper is important and timely considering the recently observed rapid increase in permafrost temperatures, documented widespread permafrost degradation, and the disappearance of permafrost in areas near the southern boundary of permafrost distribution. These processes have a profound impact on the environment, carbon cycle

I do not have any serious objection to the content of this paper and how this content is presented, however I would like to make several suggestions that may improve the paper:

1. Lines 39-40: Not clear. More explanation and a reference or references is/are needed here to explain better
2. Lines 46-47: Not clear statement. The difference between these two models is that TTOP is an equilibrium semi-analytic model, but the CryoGrid is a transient numerical one.
3. Line 88: “Ground sensitivity analysis” is a confusing terminology. Better to call it "model calibration"
4. Figure 3: In the upper right block of text: Volumetric content of what and how is it tested?
5. Line 96: “…and integrated in the PERMANTAR…”?
6. Lines 105-106: Why monthly max, not the daily averaged temperature profiles?
7. Lines 114-115: Not clear statement. Suggest more explanation
8. Line 119: What is the depth of this column and what kind of lower boundary conditions were applied/prescribed?
9. Line 133: “0.05 to 0.5 cm” Is it cm or m?
10. Equation (2): in the right side of the equation (t,z) and z should not be subscripts
11. Line 144: 0.99 plus 0.02 makes 1.01, not 1 as it should be
12. Lines 161-162: In this case is it logically to assume that the rest of the parameters (from a) through g)) will be also significantly deviated from the real ones? How can you correct this bias?
13. Lines 178-179: “averaging 3.3 C for the season” – Is it air temperature or ground surface temperature? Cold season or mean annual?
14. Line 181: “A similar pattern…” - What do you mean under "similar pattern"? Ground surface temperature during TS was higher in 2020/2021 than in 2021/2022, but it was colder for FS of 2020 compared to 2021 and 2022 freezing season
15. Figure 5: Will be good to add a 0 C line to this Figure. It will help to understand the freezing/thawing process and the timing of it much easier
16. Figures 6 through 8: Very difficult to read numbers on the graphs. Suggest increasing the fonts
17. Lines 257-258: I cannot see any evidence of the "zero curtain effect" in Figure 5. This Figure shows that the temperature in the active layer crossing the 0 C temperature threshold practically simultaneously at all depths. In this case we cannot talk about "zero curtain effect".
18. Lines 278-279: Please, provide here or in Discussion section some explanation of this effect and why the model overestimates ALT
19. Figure 9: How was this figure produced? It seems that it reflects the freezing process of the active layer not realistically. It shows that the freeze-up is happening form the bottom up. In contrast, Figure 5 shows that the temperature at the all observed depths in the active layer crosses the 0 C threshold practically at the same time
20. Line 289: “Severity” is a strange choice of wording. "Severe warming" sounds a bit strange
21. Line 290: We cannot see TDDs in Figure 10.
22. Line 311: My calculations show 0.51 C/decade for MAGST ((4.5-0.8)/7.2)
23. Line 312: Why is the geothermal gradient in the upper 20 m prescribed in the model for 1950 so large (0.09 C/m)? Also, my calculations of the warming rate at 20 m depth show 0.21 C/decade
24. Figure 12: More contrasting line colors will be very desirable
25. Line 327: Definitely wrong unit (probably should be cm or mm). Also, it needs to be said that it is snow water equivalent, not the snow fall
26. Line 328: ALT is controlled mostly by the summer conditions, not by MAGST or TTOP which are mean annual values.
27. Line 334: Please, check this number. It is not small and equal to the rate of warming during the Period 1
28. Line 347: Please, find a better word here. "Severity" of the thawing season doesn't sound good for cold climate regions
29. Line 386: This fact was not explained in the Results section. May be it is a good place to speculate what was the reason in this overestimation of ALT in the model

I can conclude that this paper can be published in the EGUsphere after minor revision.