This manuscript presents a solid study that integrates the global glacier model GloGEM with the HBV model implemented in the Raven framework, together with a snow redistribution scheme, across 14 glacierized headwater catchments in Switzerland. The study offers a systematic assessment of how glacier constraints, snow redistribution, and precipitation correction affect melt partitioning and parameter behavior at the catchment scale. Overall, the work is carefully conducted and valuable. That said, I have several concerns regarding the experimental design and the novelty of the main findings. My main comments are as follows:

1. The main contribution of the paper seems stronger as a systematic regional assessment than as a genuinely new methodological advance. The study is solid and useful, but the novelty of the main findings appears somewhat limited. The main messages of the paper are that (i) good streamflow performance does not necessarily imply realistic cryospheric process representation, (ii) uncoupled HBV can compensate forcing biases through excessive glacier melt, and (iii) introducing a precipitation correction factor can aggravate equifinality and uncertainty. These are important points, but they are not new in cold-region hydrology and glacier-hydrology modeling. In this sense, the main value of the manuscript seems to lie in the systematic multi-catchment comparison across 14 headwater basins, rather than in delivering new methodological or process-level insight.
2. The role of the precipitation correction factor (PCF) needs to be explained more clearly. The manuscript lists PRECIP_CORR as a precipitation correction factor in Table 2 and treats it as a separate configuration dimension in Table 3 (e.g., HBV-PCF and HBV-GloGEM-PCF). However, its hydrological meaning and implementation remain insufficiently explained. At present, the manuscript provides a label, but does not clearly explain what exactly is corrected, how the factor operates within the HBV/Raven setup, or what physical bias it is intended to represent. It is also unclear why calibration of this parameter is treated as a dedicated scenario rather than simply as one additional calibration parameter.
3. The contrast between the two model setups may not primarily reflect “whether GloGEM is coupled,” but rather whether glacier information is used to constrain the glacier component. In the current design, the coupled setup benefits from glacier-specific calibration against geodetic mass balance, whereas the original HBV setup is calibrated only against streamflow. This makes it difficult to attribute the differences solely to coupling with GloGEM. In practical terms, both the GloGEM-based runoff and the melt generated by the simplified HBV glacier routine ultimately act as inputs to the hydrological simulation. It is therefore plausible that, if the native HBV glacier routine were also constrained using glacier data, it could produce more comparable glacier-melt estimates. The manuscript should discuss this point more explicitly and clarify whether the key distinction is really coupling itself, or the use of glacier observations to constrain the glacier contribution.
4. The snow redistribution module should be introduced and visualized more clearly. The manuscript explains why snow redistribution is needed and briefly describes how it is implemented. However, its role is not clearly shown in the workflow figure, which makes the overall methodology harder to follow. I suggest either adding this component explicitly to Figure 2 or clarifying its position and function more clearly in the text.
5. The treatment of snow-covered area (SCA) remains unclear, and the lack of SCA-based evaluation weakens confidence in the snow and melt partitioning results. I may have overlooked it, but I could not find a clear explanation of how snow-covered area is determined in the model. In addition, the model does not appear to be evaluated against SCA data. This limits confidence in the snowmelt simulations. Even if catchment-average SWE is reproduced reasonably well, the spatial distribution of snow and the timing and location of melt may still remain uncertain without an independent constraint on snow extent. I therefore recommend that the authors clarify whether SCA is represented or diagnosed in the current workflow, and discuss the implications of not evaluating SCA for the interpretation of snow storage and melt contributions.

Minor issues

• The naming of the model configurations in Table 3 is somewhat confusing. In particular, the HBV-PCF and HBV-GloGEM-PCF configurations also include snow redistribution, but this is not reflected in their names. A more consistent naming convention, or a clearer explanation when these configurations are first introduced, would improve readability.
• In Table 2, BASEFLOW_COEFF is listed twice for the fast and slow reservoirs. Although this may be correct in the model implementation, the presentation could be clearer for readers who are not familiar with the Raven/HBV setup.
• Please clarify explicitly whether SWE was used only for validation or also during calibration. From the current text, SWE appears to have been used for validation rather than calibration. Since SWE results are shown prominently, this point should be stated unambiguously.
• There are several minor language and typographical issues that should be corrected during revision. For example, “largly” should be “largely” (line 85), “relativity similar” should be “relatively similar” (line 230). Figure 2 appears to contain a typo in the label “Precipiation per HRU”, which should be corrected.

The manuscript presents a systematic modelling experiment in which the authors couple the global glacier evolution model GloGEM with the HBV model implemented in the Raven framework, and assess the added value of glacier constraints, snow redistribution, and precipitation correction for simulating streamflow and partitioning snowmelt and glacier-melt contributions across 14 glacierized Swiss headwater catchments. The topic is relevant, and the paper addresses the fact that good streamflow performance does not necessarily imply a realistic representation of snow and glacier melt processes. Overall, I find the study valuable, but I think that some methodological choices need to be clarified before publication. The major issues in my opinion are:

Role and validation of snow redistribution: snow redistrbution appears to have an important influence on the partitioning between snowmelt and glacier melt. In particular, it may affect how long glacierized areas remain snow-covered, and therefore how much ice melt is simulated. Given this importance, the snow redistribution scheme should be described more clearly, including how it operates within the semi-distributed HRU structure. I also think that an independent evaluation would be useful, for example using maps snow-covered area from satellite observations. Catchment-average SWE alone may not be sufficient to verify whether the spatial distribution and timing of snowmelt are realistic.

Routing of GloGEM meltwater: the treatment of glacier runoff in the coupled configuration is not fully clear to me. The manuscript states that GloGEM runoff bypasses intermediate storage components and is directly added to the runoff–streamflow transfer function. However, the native HBV glacier module seems to include a glacier storage coefficient controlling meltwater release. Are glacier melt contributions routed in a comparable way in the HBV and HBV-GloGEM configurations, or does the coupling remove part of the glacier storage/routing representation? Since this may affect the timing of simulated discharge, especially during the melt season, this point should be clarified and justified.

Precipitation correction factor: the hydrological meaning and calibration of the precipitation correction factor should be explained more clearly (i.e., is this factor intended to correct precipitation undercatch, orographic effects, winter precipitation bias,etc.). Is this PCF different from the fc,prec (line 205) calibrated for the GloGEM? How do these two parameters are linked? If they are different, does this means that there is a precipitation correction in all the HBV-GloGEM coupled configuration? Would this change the volumes reaching the catcment between configurations?

Initialization of model states: the initialization and spin-up of the different configurations should be clarified. Are all configurations initialized with the same SWE and storage conditions? Is SPASS used only for validation or also for initialization? Why not for calibration? In Figure 5b, the different configurations already seem to show different SWE values at the beginning of the validation period. This initial offset should be explained, as it may affect the interpretation of SWE evolution and “snow towers” development.

“Model parameters” section: this section is very short and the entire paper could benefit from a more extended description (and discussion in the “discussion” section) of the changes in model parameters (not only the melt factor) among the different configurations and how such differences are related to model overcompensation to fit the streamflow calibration objective. I see the authors have included figures in the supplement but I would appreciate a more extended presentation in the text.