This study provides an analysis of the Crocus-ERA5 snow product covering the Northern Hemisphere from 1950-2022 through comparisons with in situ observations, satellite data, and its predecessor (Crocus-ERAI). The paper can potentially provide an important contribution; however, there are key issues that need to be resolved prior to publication. Importantly, it is essential to more carefully address the limitations associated with vegetated areas, consider observational uncertainties, and improve writing and figure quality. More details on the dataset itself are needed. Specific points are noted below.

The grammar and spelling throughout should be carefully edited.

The limitation of the Crocus-ERA5 product regarding open areas with low vegetation should be included in the abstract, as this largely alters the domain of the dataset’s applicability.

Introduction: the novelty of this work should be more clearly highlighted

Line 80 and throughout, snowpack should be a single word without spaces.

Lines 99-101: Snow cover area, rather than snow depth, plays a more dominant role in the surface energy balance. There should be a better justification for the focus on snow depth (e.g., rather than SWE). If there is not a strong reason why SWE could not be evaluated, then an evaluation of SWE should be included.

Line 122 and throughout: “extent” instead of “extend”?

Line 123: “climatic warming” rather than “cooling”? Reduced snow cover extent decreases surface albedo, which in turn favors enhanced warming. Also, the snow-albedo feedback should be discussed in this paragraph as an important motivator for evaluating snow cover extent.

Line 124: if this study/data is not applicable across boreal forests, perhaps this should be removed.

 Section 2: There should be a paragraph dedicated to observational details (resolution, methods, uncertainties, scale mismatches, etc.)

Line 158: Consider using a word other than correlated (i.e., “related”), because correlated assumes a linear relationship, whereas the snow depletion curve is non-linear.

Line 164: consider using “ablation” here because sublimation and play an important impact in some areas.

Methodological additions recommended: Please include more detail on the methodology used by the model to estimate snow cover extent, snow albedo, and the role of snow cover in the model in estimating the land surface albedo. Please include more detail on how vegetated areas were screened in the model evaluation procedure.

However, figures showing spatial distributions do not look to screen out vegetated areas, despite this model limitation. Lack of screening can introduce a large amount of noise in spatially averaged comparisons, and thus screening should account for this important dataset limitation. If you wish to show the bias attributable to the lack of vegetation representation, please show results both with and without vegetation screening.

Figure 1 should also show anomaly difference plots. It would be useful to note the mean snow depth that is being subtracted from the respective datasets.

Line 181-182: this seems counter to Figure 1 panel 2, which shows above average snow depth in July and August for all years. Likewise, panels 3 and 4 show above average snow depth during many summer months. Is this erroneous? Considering the description of the annual cycle here, perhaps Fig 1 should be ordered by water year, rather than calendar year.

Figure 2: “snct” is not defined. Please update the y-labels in these figures. The x-label title “Year” is not required. Similar to the comment above, I think it makes more sense to show these data by water year, rather than calendar year. Why show the data using standardized anomalies here which has a less physical meaning? Please be clearer in what the solid lines are vs. what the circles are. Why only show Apr-Jun? These comparissons also seem important for accumulation period months. There may also be a better way to show the comparisons, e.g., showing ERAI and ERA5 on the same panels, and separating panels by NA and Eurasia.

Figure 3: labels need to be updated in this figure. In the corresponding main text (lines 169-175) there should be a few sentences dedicated to what the comparison of the stdev vs. average is indicative of, and why this comparison is chosen.

Figure 4: a difference panel would be useful.

Line 193 and throughout: “0.75” rather than “0,75”

Figure 5 labels should be updated. For instance, “Canada” panel also shows Alaska and Greenland. The caption should fully describe the figure. Time series panels need a legend. Lower xlimit should start at the beginning of the time series comparisons.

Figure 6: “continuous”.  Quality needs improvement.

Figure 7 would be more insightful if snow cover faction, rather than anomaly is presented, and perhaps a panel showing the difference. It is important to present the systematic biases as well.

Figs 7-8: Vegetation screening should be considered in these evaluations.

Figure 9: where is the comparison noted in the caption? Are these difference maps? Please specify. Highly vegetated areas should be screened out. Oceans and screened areas should be white, rather than a color that corresponds with the color bar.

Line 272-280:  The methodological section should note which statistical tests are used in this study. The p-value noted here should include a specification of which test this corresponds with. Standardizing the time series would force an identical mean and standard deviation between the 2 compared time series, perhaps the statistical test should be used on the raw, rather than standardized, time series. The trend should be based on raw data, rather than the anomaly time series. The MAE between the two curves should be 0, because the anomaly timeseries shown in Figure 8 should have means = 0. If the MAE is based on the raw data, this should be specified.

The dataset contains SWE, snow albedo, surface temperature, snowpack temperature which are not evaluated. A quality flag for the dataset, pertaining to heavily-vegetated areas, is recommended.