Review of Védrine et al.'s: "Temperature sensitivity of snow viscoplasticity: evidence from controlled creep experiments"

General comments on the manuscript:

The manuscript by Védrine et al. presents an experimental investigation of the temperature dependence of snow viscoplasticity using a newly developed thermo-mechanical cell coupled with in situ X-ray tomography. The experimental protocol, in which temperature is systematically varied on the same snow sample while monitoring microstructural evolution, represents a clear improvement over previous approaches that rely on separate samples at different temperatures and initial bulk densities. The resulting measurements provide insight into the temperature sensitivity of snow deformation and have important implications for constitutive models of snow settlement and snowpack evolution.

Specific comments for the authors' consideration:

1. The identification of two activation-energy regimes is one of the central conclusions of the manuscript. While the data are certainly consistent with a bi-Arrhenius relationship, Section 4.3 also demonstrates that the observations can be explained by introducing a temperature-dependent stress exponent. I wonder whether the authors might consider framing the proposed transition near −13 °C somewhat more cautiously, emphasizing that the data are consistent with—but do not uniquely require—a distinct physical transition.
2. I wonder whether the discussion in Section 4.3 could also briefly consider the possible contribution of grain-boundary sliding to the observed temperature sensitivity. While the authors discuss basal slip as the dominant deformation mechanism based on Védrine et al. (2024), it seems worthwhile acknowledging whether grain-boundary sliding might also contribute to the mechanical response, particularly given the fine-grained, porous nature of snow.

Overall assessment:

Overall, the manuscript is well written, the experiments are well designed, and the conclusions are generally supported by the presented data. The manuscript presents a novel and technically strong experimental study that substantially improves our ability to isolate the influence of temperature on snow viscoplasticity. The newly developed thermo-mechanical cell and experimental protocol represent an important methodological advance, and the resulting observations provide useful constraints for the future development of constitutive models of snow deformation. Subject to minor revision, I believe this manuscript is well suited for publication in The Cryosphere.

Editorial comments keyed to line numbers:

• 282 – Change "a activation energy" to "an activation energy."
• General – Several figure axis labels, legends, and mathematical symbols appear corrupted in the review PDF, likely due to a font-embedding or PDF rendering issue. I recommend verifying that the figures are exported with all fonts properly embedded before publication.

The viscoplastic creep of snow and ice is highly temperature-sensitive, following an Arrhenius-type relationship where strain rate increases exponentially with increasing temperature. The activation energy term governing this temperature dependence remains among the most uncertain parameters in constitutive formulations due to the difficulty of isolating thermal effects during deformation from the simultaneous microstructural evolution, i.e., grain growth, recrystallization, geometric hardening from densification.

The manuscript by Vérdine et al. provides novel experimental constraints on the values of apparent activation energy for snow viscoplasticity with a newly developed thermo-mechanical cell. This methodology improves on previous approaches as it (1) eliminates the variations of the initial microstructure and bulk density by varying the temperature of the same sample and (2) accounts for the concurrent microstructural evolution through in situ X-ray tomography coupled with a viscoplastic model previously developed by the authors in Vérdine & Hagen Muller, 2025.

Overall, the manuscript is well-written, the experiments presented therein are novel and rigorous, and the reported apparent activation energies may be readily applied to snowpack modeling. The conclusions reached are reasonably supported by the data but would benefit from an expanded discussion on the micro-scale origin of the presented two-regime Arrhenius relationship. The manuscript is well suited for publication in The Cryosphere but would benefit from some minor revisions.

Comments:

1. The activation energy is fundamentally the intrinsic energy barrier associated with a single thermally activated mechanism. However, deriving this quantity from the temperature dependence of the strain rate using an Arrhenius plot is only justified if the measured deformation is controlled by a single mechanism over the temperature range considered. Given the potential of enhanced recovery or the influence of grain-boundary processes especially at higher temperatures, I do not believe this can be stated with certainty from the data presented in this manuscript. I would therefore suggest referring to the fitted Q values reported in section 3.2 as an apparent activation energy, and elsewhere throughout the manuscript as pertinent.
2. The bi-linear fit reported in Section 3.2 appears to be in respectable agreement with the mechanical data and occurs at a transition temperature that is generally consistent with that observed in dense polycrystalline ice. In Section 4.3, the authors argue that this behavior may not reflect a true transition in regime but could instead arise from enhanced basal slip activity in the ice matrix. Supporting this interpretation, experiments in which single crystals of ice favorably oriented for basal slip were deformed at various temperatures, report an increase in the stress exponent, n, with increasing temperature from -20 °C to -0.2 °C (Jones and Brunet, 1978). When this stress exponent temperature sensitivity is considered (n(T) = 2.15 – 0.00726 × T), the bilinear trend disappears, and the scaled strain rate prefactor becomes quasi-linear with respect to the temperature on the Arrhenius plot shown in Figure A.3. For the author’s consideration, I have some specific comments concerning this treatment:

• It is unclear to me how the equation describing the temperature sensitivity of the stress exponent is obtained in section 4.3. A quick fit to the data from Jones and Brunet, 1978 returns n(T) = 1.95 – 0.0061× T which differs slightly from the equation presented (n(T) = 2.15 – 0.00726 × T). I assume that this equation may have been normalized to the mechanical data in some way to ensure that the fit passes through a stress exponent of n = 2.2 at a given temperature, but this should be stated more explicitly in the manuscript for clarity particularly given the difference in slope between the two equations.
• The linear trend of the temperature sensitivity of the stress exponent persists from -20 °C to -0.2 °C, but the mechanical data reported in this manuscript covers a temperature range of -6 °C to -18 °C. In polycrystalline ice, there is evidence that the creep rate accelerates above T = -5 °C as the temperature approaches the melting point beyond what would be predicted from a linear fit to the -5 °C to -10 °C behavior (e.g. Russel-Head and Budd, 1979, Jacka and Budd, 1989, Morgan, 1991). This may be present in snow as well, manifesting as apparent activation energies ranging between 201 and 426 kJ mol^-1 above -5 °C reported in Scapozza and Bartelt, 2017 (though they are unable to isolate the microstructural effect without in-situ x-ray tomography). This suggests the creep rate, especially above -5 °C, may appear non-linear on an Arrhenius plot approaching the melting point, and it is unlikely that the linear change in the stress exponent that the authors attribute to basal slip activity will fully capture the change in activation energy approaching the melting point. Section 4.3 would benefit from an expanded discussion to include the potential role of intercrystallite mechanisms such as grain-boundary sliding in the high temperature regime and encourage the readers to exorcise caution extrapolating the apparent activation energies above the experimental range presented here.
• One of the limitations of the treatment of the stress exponent derived in the Jones and Brunet, 1978 is the use of yield stress, corresponding to the stress at which dislocations become mobile and multiply, as a fitting parameter which may not be appropriate to apply to the creep behavior; necessitating some additional justification from the authors.