General comments:

Pletzer at al. present a point-based optimization and evaluation of a simplified version of the WRF-Hydro/Glacier modelling system over the McMurdo Dry Valleys (MDVs) using forcing data from automatic weather stations. They identify two aspects of the cryospheric component needing improvement for accurate simulations of runoff, namely the representation of percolation in ice layers and the parameters in the albedo scheme. The manuscript is well-written and organized, the results are clearly and concisely presented, and the topic suits the scope of the journal. However, I have a concern about the novelty and wider applicability of the presented results due to the simplified model configuration employed, as outlined in my major comment below, which should be considered prior to publication.

Major comment:

As I understood, the authors only used the cryospheric (Crocus) and land-surface (Noah-MP) modeling components from WRF-Hydro/Glacier. The presented work is therefore mainly small changes to/calibration of parts of Crocus, which is an important foundational step for tackling interesting science questions in a new region but is itself a methodological task. As a result, the introduction lacks a clear scientific question and, in my opinion, the results may be insufficiently novel, as previous studies have applied Crocus in Antarctica (including in coupled simulations; e.g., Brun et al. (2017) https://doi.org/10.3189/002214311797409794) and performed point surface energy/mass simulations in the MDVs (line 383).

The authors argue for the standalone approach to “limit uncertainties in meteorological forcing data introduced when coupled to WRF” (line 144), however observational data also contain uncertainties (e.g., the discussion around deriving solid precipitation from SR50s). More importantly, this simplification means that the capability of the full WRF-Hydro/Glacier modelling system has not been assessed and leaves open the question of how the presented modifications impact simulated runoff in coupled simulations when changes in surface conditions can feedback on the atmospheric forcing.

I suggest that the authors either provide a stronger justification for their approach and/or more clearly communicate the novelty and advancement in scientific understanding of their work, or ideally include additional experiments with the full WRF-Hydro/Glacier model (for example, comparing oldrunoff_oldalbedo and newrunoff_newalbedo in an interactive context). The latter suggestion would greatly strengthen the impact and novelty of the manuscript.

Specific comments:

1. The fact that the manuscript presents a standalone simulation without the atmospheric or streamflow components should be more consistently communicated. Although the AWS forcing is mentioned in the abstract, this simplification is not clear from the title or from using the full model name throughout the paper. Statements like „the first application of WRF-Hydro/Glacier model in MDV“ (line 5) while technically correct also do not communicate the nuances of the presented work.



2. Please provide information about the extent and location of the 200-m computational grid and how the point AWS data was distributed spatially over this grid. However, it would be worth mentioning why this grid was used when the two active components (Crocus and Noah-MP) are column models without any lateral interactions and the analysis is point-based.



3. Please provide information on time periods for spin-up vs simulation, calibration vs evaluation, and CWG vs COHM forcing earlier in the methods and in one place for convenience (e.g., in a table).



4. Section 2.1: Please provide more detail about what running in standalone mode means in terms of active components and/or interactions.



5. Line 150: Four months seems short for spinning up ice temperatures in a 50 m column based on my experience with the heat equation. Which objective criteria were used to determine if this period was sufficient? In Figure 8, it seems the biases change systematically in time (i.e., the cold bias at depth decreases while at upper levels it increases). Could this feature be attributable to spin up (which could be assessed with sensitivity runs) or e.g., the changing depth of the temperature sensors?



6. How was the snow depth and density profile initialized? How well are snowpack conditions represented at the end of the spin-up period and how might this influence the simulated albedo?



7. Line 276: how was the absence of overfitting assessed?



Suggestions for technical corrections:

1. Line 14: Remove „of melt“

2. Table 1: Is it relevant to provide detailed information on instrumentation for COHM as well?

3. Line 214: Please clarify that you are referring to the Crocus snow albedo scheme

4. Line 300: How do you define slight? The cold bias is ~2 K

5. Line 307: „a cold bias and a phase shift“

6. Line 328: Rephrase „similar“ as surface height changes range approximately an order of magnitude

7. Line 350: Rephrase „ephemeral,“ as the term may not be accessible to a wide audience

The manuscript submitted by Pletzer et al., is a preliminary step toward applying WRF-Hydro/Glacier to a cold-based Antarctic glacier. Generally, the manuscript is well written, figures and data are presented well and the subject matter is a good fit for the journal. There is potential for this approach to reveal new understanding of the MDV hydrologic system as a whole. However, in agreement with the other reviewer, this present manuscript represents in intermediate methodological step and demonstrates no real advance in scientific understanding as written.

Major Comments:

As this work is presented, the introduction suggests that the full WRF-Hydro/Glacier model is required to make inferences about glacier-stream-lake hydrologic connectivity. However, the rest of the manuscript is focused on the point-based simulation and tuning of Crocus and NoahMP rather than the full model itself. There is no scientific question or hypothesis to be addressed. This paper at a minimum should be reframed to identify and explicitly address the scientific advancement made by this paper alone, not the future applications of the model. A more impactful contribution could involve running experiments with the model.

Minor Comments:

Why was such a short period used for model spin-up? Since the COHM met station was used for spin-up anyway, there are many years of data available from that site. Why limit it to a few months? Also, as the other reviewer notes, might this have some impact on temperature bias shown in the results?

I’d like to see more discussion on how the 2021-22 season relates to the long-term average climate here. Was this a warm, cold, snowy, cloudy etc. season? Since the modified albedo scheme (and overall model tuning and results) were so dependent on data from a single season, it would be nice to provide more context for this season relates to typical summer conditions for this glacier. There is potential to overfit the model for this set of conditions and it may not perform well for colder or warmer seasons. Was this assessed?

Please provide more detail on the vertical layer thickness and spacing. Based on figure 3, it is not uniform. This scheme is critical for accurately representing processes that have strong gradients in the shallow subsurface.

How was overfitting assessed for the Albedo modifications?

Please provide more detail and comparison of data, instrumentation and accuracy across the two met stations.

L121 – Sentence is a repeat from on the previous section

Section 4.2 – The beginning section as well as a few introductory sentences in paragraphs elsewhere in the section are basically only telling the reader what they will be told later and are therefore unnecessary and should be rewritten.

L355 – this citation is incorrect.