This paper employed an S3M model to blend multi-source in situ data and satellite observations to produce a spatially explicit and multi-year reanalysis of snow cover patterns across Italy at 500 m resolution. After validating with C-SNOW products, in situ measurements, and annual streamflow, this product has been proved effective, and could be potentially used in better understanding the contribution of snow on water resource management.

Despite of its significance, several issues still need to be resolved before a publication to ESSD. More detailed introduction about how to produce snow cover area from multi-source remote sensing images, and how to produce the reliable snow depth maps over the entire study could be sufficiently explained. In addition, it is suggested to add more indexes to validate the output snow estimates. Besides, the figures should be further refined so as to improve the overall presentation.

Other comments and suggestions:

1. Figure 1, the schematic of S3M was too simple, it is difficult to understand the key model/method, the data flow, and the output data.
2. P115-116, please provide the elevation gradient for air temperature when you interpolated in situ air temperature.
3. P125-135, the snow covered area used in S3M model are produced from Sentinel 2, MODIS, and H-SAF initiatives. How did you produced snow cover area from Sentinel 2? How did you preprocess the MODIS and H-SAF data? Have you filled up the data gaps caused by cloud cover? How to fill the data gaps? How about the accuracy of the blended snow cover area products?
4. Figure 2-4, and 8, please add scale bar and change the color of latitude and longitude grids from black to white or gray. It is difficult to identify detailed numerical value from current stretch effect of color bar.
5. P145-155, the in situ stations are primarily distributed in north areas in Figure 2 (a), so how did you produce the reliable snow depth maps over the entire study area? How about the overall accuracy of the daily snow depth maps over the entire 10 homogeneous regions? If some of the homogeneous regions are lack of snow depth data, how about the final output after running the S3M model for these regions?
6. Figure 3, please add legend for (a); it is cannot see NaN class (in orange color) from (b); add scale bar for (a)-(c).
7. Figure 4, why did not show the results over the entire study area?
8. For the validation results, please also add Mean Absolute Error, Positive Mean Error, Negative Mean Error, and R Squared.

This paper discussed a new spatially distributed Italian snow reanalysis through combining remote and in-situ measurement techniques with the already existing Snow Multidata Mapping and Modeling system (S3M). Evaluation of the reanalysis through comparison with separate in-situ (snow course) and remote sensing products (C-SNOW) showed reasonable error within the produced snow products including snow depth, snow water equivalent, and snow density. The output products showed agreement with inter- and intra-annual accumulation and ablation trends in various climatological regions throughout Italy where different snowpack characteristics exist.

The reanalysis and associated paper(s) are novel and show significant potential for use with climatological analysis and monitoring of the Italian snowpack, and the overall grammar and organization of the manuscript were good with minimal issues. However, revisions are required to improve the manuscript before it should be accepted to ESSD.

Major comments:

1. It would be useful to have analysis of average error of snow depth, SWE, and density for each of the 10 homogeneous regions mentioned first on Line 146 and shown in Figure 3a. Given the distinct geographical and climatological characteristics of each region and non-uniform distribution of the in-situ sites, regional differences in error may be expected that would be important for users of this data to understand. It would also aid in the constraint of the relative importance of SWE in each of the basins discussed in Section 4.2 and Figure 10.
2. Lines 115-116: Further information about these linear regressions should be presented. How were they derived and applied?
3. Line 162-163 and Figure 3: It is discussed that SCA maps are not assimilated but are used to clip pixels that are snow free from snow depth maps. Figure 3 shows the SCA and snow depth maps individually but it would be helpful to have an additional panel showing the post-SCA clipped snow depth map to highlight the data that is being assimilated.

Minor Comments:

1. Line 1: “The” at the beginning of the sentence can be omitted.
2. Figure 1. Further detail is needed in this flowchart. Specific information on the meteorological variables as discussed on Line 101 as inputs should be displayed.
3. Line 147-148: “expert knowledge”. This doesn’t need to be exhaustive, but it would be nice to know what other primary conditions were considered in the expert knowledge.
4. Lines 233-237 and Figure 4d: Distribution of root mean squared error in Figure 4d shows a right skew. As such, median should be used instead of mean.
5. Line 284 and Figure 8b: Same as above. Data shows right skew and median should be used rather than mean.
6. Line 353: Should be ‘1st’, not ‘1th’.
7. Figure 3a: Add legend.
8. Figures 2, 3, 4, 5, 6, 7, 9, and 10: Color blind-friendly color palettes should be implemented.
9. Figures 2, 3, 4, and 8: Can’t see lat/long grid lines. Suggest changing to more visible color.
10. Figures 2, 3, 4 and 9: Increase size of color bars/scales and add additional values.

Review of "IT-SNOW: a snow reanalysis for Italy blending modeling, in-situ data, and satellite observations (2010-2021)". Avanzi et al.

This article presents a reanalysis of the snowpack conditions over the italian territory between 2011 and 2021. It uses a spatially distributed snowpack model (S3M) forced with gridded in-situ observations from automatic weather stations (AWS) and radar. The simulated snow depth is corrected by the assimilation of snow depth measured at AWS gridded thanks to a multilinear regression model and adjusted by satellite-based snow cover maps. The uncertainty of the reanalysis is estimated with Sentinel-1 derived snow depth (C-SNOW) and in-situ snow depth and SWE measurements.

This high-quality reanalysis will sure be useful for many applications. The article is well-written and seems comprehensive, covering most aspects of this work. The methods and results are well presented. I believe that the following points should be addressed by the authors before publication. Below are smaller suggestions and details to help improve the article.

1. The figures need improvement, especially figures 2 and 3. Each map needs a title. The axes should be labeled, a scale added. The colorbar choice often does not allow a clear reading of the maps. The colorbar legend is often too small and with too few labels. See detailed comments on each figure below.

2. Some data and methods information seems missing. I could not find which digital elevation model is used (what source, what resolution) or if the land cover is taken into account. It would be good to mention if the interaction between the snowpack and the vegetation, such as the forests, are considered.

3. Other reanalysis over the swiss, the austrian and the french Alps (Fiddes et al., 2019, Olefs et al., 2020, Vernay et al., 2021) are mentionned. Although the methods are largely different in each work, it would be interesting to compare the uncertainty of these works.

Minor comments and suggestions

L2 "+" disturbing notation. I suggest using "over", ">" or just give the exact value. To be homogeneized in the text.

L9 "no mean bias" rather than "none"? (L421 as well)

L14 If ever the variability of the peak SWE date is available, it could be interesting to provide it.

L25 "(Serreze et al., 1999; Skiles et al., 2018)" you might want to cite Li et al. (2017) in which the contribution of the snowpack to the runoff is indeed calculated. It seems like Serreze et al. (1999) only compared the solid precipitation amount to the total runoff and Skiles et al. (2018) cites Bardsley et al. (2013) for the 80% number.

How much runoff originates as snow in the western United States, and how will that change in the future? Dongyue Li, Melissa L. Wrzesien, Michael Durand, Jennifer Adam, Dennis P. Lettenmaier. GRL. 2017, https://doi.org/10.1002/2017GL073551

L38 "lidar" in Deems et al. (2013), "Lidar measurement of snow depth : a review". To correct everywhere.

L38 "airborne lidar"? otherwise the list mixes methods (lidar, optic) and plateform (drone, satellite).

L47 and further in the text: what is a "dynamic model"?

L67 GlobSnow: maybe worth to mention that it is not available in mountain areas?

L91. A bit confusing with S3M, S3M Italy and IT-SNOW. Maybe add "the reanalysis IT-SNOW"

L100 "**" => I was disturbed by this notation without letters. Maybe use "hh" instead?

L100 Maybe precise the period covered by the inputs: is it only of the last hour?

L108 RMSE of 1 mm, please provide the typical precipitation observed.

L112 "spatialized" at what resolution?

L115 It would be very useful to provide the distribution of the temperature lapse-rate, even if supplement in necessary. This study from Navaro-Serrano et al. (2018) might help if you need to compare your temperature lapse-rate to similar regions.

Navarro-Serrano, F, López-Moreno, JI, Azorin-Molina, C, et al. Estimation of near-surface air temperature lapse rates over continental Spain and its mountain areas. Int J Climatol. 2018; 38: 3233– 3249. https://doi.org/10.1002/joc.5497

L118 I would suggest rewording along "An unique estimate of the precision of these data is not available as the type of sensor installed varies from one region to another. The installation and the maintenance of the sensors..."

L122 "remapped" quite vague. Cropped?

L124 "each region to tailor" unclear. What is the exact meaning of "region" here? What is tailoring S3M?

L128 "Sentinel-2"

L129 How do you manage the overlapping images? Putting on top the most recent?

L137. "Not shown". Could be added in supplement maybe?

L146 Please provide the number of snow depth sensor

L151 "remapped"? unclear.

L159. "For each time instant" not clear. Could be deleted.

L163 What happends if snow in SCA observation but not in S3M? "preserving" is a bit unclear, maybe use "leaving without snow..."?

L170 "The duration"?

L171 "1.3 h" give it in h and min.

L172 "AM" a.m.?

L184 Given the resolution, it seems like at least the last "57" can be dropped.

L194 Some precisions about C-SNOW product would be welcome. First, it is only available for dry snow, that is accumulation period, isn’t it? Second, some part of Italy seems not covered by C-SNOW (grey area in Fig. 1 of Lievens et al., 2019). Finally, is C-SNOW completely independant from IT-SNOW? C-SNOW was calibrated on snow depth from AWS.

L211 "ASL" a.s.l.?

L232 Is it not possible to make it "3.2 Results" and then sub-sections (3.2.n) for the different data sources?

L235 how did you compute the RMSE? Between time-series at each pixel? Please write the MAD from Lievens et al. (2019).

L243 Please provide values for the bias. A table summing all the statistics evaluation would be really helpful.

L248 Then there is little information brought by the comparison of IT-SNOW density with station density since the density is derived from snow depth and SWE and snow depth and SWE are also compared to IT-SNOW.

L253 Please provide bias values.

L284 "102 gauge stations"?

L294 "Again,..."

L306 "evalution results"=> "the results"

L307-309 Cut this long sentence int two.

L313 "peripheral"? geographicaly peripheral? The Alps are on the periphery of Italy but the station density is high...not clear.

L317 not clear if talking about the SCA of the Sentinel-2/MODIS product or from IT-SNOW.

L331-334 sentence is too long. To be cut.

L334 "apriori appear"? please rephrase.

L336 "Thus, quantifying this uncertainty in still elusive at this stage." I dont understand where this conclusion stems from.

L336: "in"=>is

L337 To move earlier in the description of the data or method.

L344 "basic science"? Please reformulate.

L348 I like the catchy quesions. However, "what is it doing?" is not so clear and it does not appear in the conclusion. Could it be deleted? I also suggest more detailed formulation "How much is accumulated in total? Where/how is it spatially distributed?"

L353 1st

L353 "(this finding is in agreement with a recent reconsideration of this conventional date, see Montoya et al., 2014)" get this sentence out of the brackets

L364 "anecdotal data" unclear if they are data from IT-SNOW or non-scientific data.

L366 "150+ cm of fresh snow in 24 hours" give the date of this event.

Figure 2. Provide title in the figure for each subplot, next to (a,b,c). Provide more values for the colorbar of b and c. In a, is there no station with several type of measurements? Or are they hidden because points overlapp?

Make b colorbar symetric so that 0° is linked to the yellow color. In b, keep the same precision (14,8//-16,77) and meaningful values (-10,0,10…).

In the legend: "by S3M Italy and thus the IT-SNOW reanalysis" => "by S3M Italy to produce IT-SNOW." Confusing otherwise.

Scale is missing as well as xlabel and ylabel.

If you find a way to make all subplots fits in only one line, it would make better use of the space.

Figure 3 See comments for Figure 2 that can be applied here. For b, why is there a transparent area without data which is not of the colour of the NaN provided in the legend.

Figure 4 This figure is much more readable. Add xlabel, ylabel to the maps and make sure color scale is symetric centered on 0. Improve the colorbar (see above).

Figure 5 Suggestion for future figures: b and c would be better plotted with a heat-map or at least some transparency of the points.

Figure 8 Make the color scale continuous, it is really hard to read the map otherwise.