General Comments

Cao and Gruber investigate the performance of five modern reanalyses (JRA-3Q, ERA5, MERRA-2, JRA-55 and NCEP2) over cold regions, with a focus on air temperature and snow water equivalent (SWE). They show that the ensemble spread in mean annual air temperature (MAAT) in reanalyses is up 90% greater over cold regions (defined as regions with a MAAT <0^oC) relative to regions with a MAAT ≥ 0^oC. The study explicitly shows the relationship between station density and ensemble spread for both SWE and MAAT and is able to show that the reduced reanalysis performance is at least partially related to the low station density over cold regions.

The bulk of the conclusions come from Figure 1, which show the average ensemble spread in MAAT (Panel A) and SWE (Panel B) binned by MAAT. The station density in each MAAT bin, and the proportion of the grid cells within the MAAT bin covered by ice sheets and glaciers, snow cover, permafrost, and seasonally frozen ground is also shown. Herrington et al. (2024) showed a similar plot for soil temperature, that identified the average reanalysis spread in soil temperatures binned by MAAT, against sample size, though it didn’t explicitly consider station density, or the proportion of the grid cells covered by the cryosphere – so this is a novel analysis (along with the focus on SWE and MAAT).  

While Figure 1 clearly shows a clear correlation between ensemble spread, station density, and the presence of cryospheric elements, there is no attempt to separate the contributions of low station density from those related to inadequate representation of cold region processes in reanalyses. As a reader it raises questions as to what the relative contributions of station density, and inadequate process representation to the ensemble spread in MAAT and SWE are?

To me, the novel science is in quantifying what proportion of the uncertainty can be attributed to station density, and what proportion is related to inadequate representation of cold region processes, which have been thoroughly discussed in the literature (e.g. Broxton et al., 2016; Cao et al., 2020, 2022; Hu et al., 2019; Mortimer et al., 2020; Wang et al., 2019).  Thus, I recommend that the authors extend their analysis to explicitly quantify what proportion of the uncertainties or spread in MAAT and SWE can be attributed to the low station density in cold regions, and what proportion can be attributed to the inadequate process representation in the products. This will greatly enhance the contribution of the paper to the literature and provide the community with useful and quantifiable estimates of the uncertainty attribution.

Specific Comments

P2, L31: Why was NCEP2 investigated over NCEP CFSR/CFSv2, for example? NCEP CFSR/CFSv2 is available at a much higher resolution than NCEP2, and is available over the period of analysis (1991-2020).

P2, L36-L37: What do the authors mean by the statement “the decades 1991-2020 were used, likely a period of high quality for reanalyses.” Is there a particular standard by which the authors determined this? Some clarification would be helpful.

P2, L38: What is the CDS? It appears that CDS is in brackets, but the acronym is not defined. I presume this may be the Climate Data Store?

P4, L101: The all 5-reanalysis value for MAAT spread was (1.5^oC, 0.5^oC-3.0^oC) - is this statistically different from the value for the 4DVar reanalyses reported here?

P10, Figure 1: What do the dashed lines represent in Figure 1? I don’t see a dashed line in the figure legend?

P11, Figure 2: It may be helpful to have a "difference" panel between the All 5 reanalyses and the three 4DVar reanalyses to highlight the regions with the largest differences,  particularly for SWE, since it is a little harder to notice the differences in Panel B.

Technical Comments

P2, L38: replace “for” in “for International Earth Science Information Network…” with “from”.

P2, L81: replace “were” in the sentence “were slightly more permafrost…” with “where”.

References

Broxton, P. D., Zeng, X., and Dawson, N.: Why Do Global Reanalyses and Land Data Assimilation Products Underestimate Snow Water Equivalent?, J. Hydrometeorol., 17, 2743–2761, https://doi.org/10.1175/JHM-D-16-0056.1, 2016.

Cao, B., Gruber, S., Zheng, D., and Li, X.: The ERA5-Land soil temperature bias in permafrost regions, The Cryosphere, 14, 2581–2595, https://doi.org/10.5194/tc-14-2581-2020, 2020.

Cao, B., Arduini, G., and Zsoter, E.: Brief communication: Improving ERA5-Land soil temperature in permafrost regions using an optimized multi-layer snow scheme, The Cryosphere, 16, 2701–2708, https://doi.org/10.5194/tc-16-2701-2022, 2022.

Herrington, T. C., Fletcher, C. G., and Kropp, H.: Validation of Pan-Arctic Soil Temperatures in Modern Reanalysis and Data Assimilation Systems, The Cryosphere, 18, 1835–1861, https://doi.org/10.5194/tc-18-1835-2024, 2024.

Hu, G., Zhao, L., Li, R., Wu, X., Wu, T., Xie, C., Zhu, X., and Su, Y.: Variations in soil temperature from 1980 to 2015 in permafrost regions on the Qinghai-Tibetan Plateau based on observed and reanalysis products, Geoderma, 337, 893–905, https://doi.org/10.1016/j.geoderma.2018.10.044, 2019.

Mortimer, C., Mudryk, L., Derksen, C., Luojus, K., Brown, R., Kelly, R., and Tedesco, M.: Evaluation of long-term Northern Hemisphere snow water equivalent products, The Cryosphere, 14, 1579–1594, https://doi.org/10.5194/tc-14-1579-2020, 2020.

Wang, C., Graham, R. M., Wang, K., Gerland, S., and Granskog, M. A.: Comparison of ERA5 and ERA-Interim near-surface air temperature, snowfall and precipitation over Arctic sea ice: effects on sea ice thermodynamics and evolution, The Cryosphere, 13, 1661–1679, https://doi.org/10.5194/tc-13-1661-2019, 2019.