This is a solid study that makes use of new detailed high vertical resolution snow thermal conductivity measurements to assess the ability of the Community Land Model to represent the snow thermal conductivity. The authors find, somewhat unsurprisingly, that the lack of a process representation of depth hoar formation in CLM leads to much higher snowpack effective thermal conductivities than are observed at these sites. The paper includes a good discussion of paths forward, including introducing an adjustment factor and utilizing alternative snow thermal conductivity formulations. Overall, I think the authors did a thorough job responding to the comments of the prior reviewers. From my perspective this paper is suitable for publication with minor revisions suggested below.
1. I found typos throughout the revised manuscript. Many of them looked like they arose through accepting track changes wherein periods or other punctuation was lost. I suggest that the authors carefully review the manuscript to correct these typos.
2. The discussion of the snow model in CLM5 is reasonable, but there is no mention of fresh snow density. The formulation of fresh snow density was modified substantially with a new temperature relationship and the addition of wind effects. In other assessment of CLM, we found that these changes resulted in higher and more realistic bulk snow densities (van Kampenhout et al., 2017, Lawrence et al. 2019), though the lack of widespread availability of snow density data makes these assessments more qualitative than quantitative. In any case, the fresh snow density parameterization should be introduced as well, I think, for the sake of completeness in the description of the snow model. I don’t think the full set of equations need to be included, but a description would be helpful. The equations can be found here: https://escomp.github.io/ctsm-docs/versions/release-clm5.0/html/tech_note/Snow_Hydrology/CLM50_Tech_Note_Snow_Hydrology.html#ice-content
3. In a response to a reviewer, the authors note that they are exploring what could be done in terms of changes that would improve the model at the global scale. But, there is no mention of this in the text. As an interested reader, I was very curious about what approach could be taken at global scale, since this bias correction parameter would be difficult to define globally and the Sturm parameterization only really applies for tundra snowpacks. I think it would be helpful for a reader to elaborate just a bit more on the challenges here and to note explicitly that it is under investigation in an ongoing study.
4. If I am understanding correctly (I might not be, in which case the text needs to be modified to make this clearer), the same bias correction factor is applied to all the layers in the simulated snowpack. But, the main limitation of the model is that there is not representation of depth hoar that forms at the base of the snowpack. Would it be possible or advisable to apply a correction factor to just the deepest model snow layer to try to replicate this real world behavior? One reason I ask is that in a study that was just published, we found that winter daily temperature variability is substantially improved in CESM2, where improved is towards less temperature variability. We inferred in that study that this reduction in variability can be largely attributed to the higher snow densities and relatedly higher thermal conductivities that acted to reduce air temperature variability due to the larger effective surface heat sink. The results of your present study imply that the thermal conductivities in CLM5 are too high, thus possibly countering the result of our study on temperature variability. However, if the uppermost snow layers are still actually relatively dense and it is just the deepest snow layer that has low thermal conductivity, it’s possible that this result on temperature variability is still valid. Anyway, not sure that you need to address this comment much or at all in the manuscript, though it might be interesting as a point of reference to mention this paper (it’s another argument for why snow thermal conductivity matters, in addition to the soil respiration that you already mention). The paper is Simpson et al., 2022, JAMES, https://doi.org/10.1029/2021MS002880
5. The Figure 5 figure caption needs more info about what the colors mean. I was able to infer that they indicated the snow layers (nice way to look at the evolution of the snowpack layers in CLM!), but I guess some readers may not be able to recognize that immediately. |