| 24 Sep 2026
White Christmas in Germany? A High-Resolution Snow Climatology since 1950 (SNOWRAS v1.0)
Abstract. Snow cover is highly sensitive to changes in temperature and precipitation, making it an important indicator of climate variability and climate change. It also strongly influences hydrological processes, for example, groundwater recharge, runoff formation, and flood risk during snowmelt. Long-term, spatially detailed information on snow depth and snow cover duration is therefore needed for climate monitoring, hydrological applications, and impact assessments.
Here, we present SNOWRAS, a new gridded snow depth dataset for Germany covering extended winter seasons from 1950/51 to 2025/26. The dataset is based on daily snow depth observations from the extensive monitoring network of the Deutscher Wetterdienst (DWD), complemented by measurements from partner networks in Germany and neighboring countries. After quality control and data cleaning, the station observations were interpolated to a regular 1 × 1 km grid using an optimal interpolation scheme.
Based on this dataset, we derived and analyzed several snow-climatological indicators, including seasonal mean and maximum snow depth, the number of snow days, and the timing of the first and last days with seasonal snow cover. Across Germany, most indicators show a statistically significant decline over the past 75 years. The magnitude of these changes varies considerably with elevation: the largest decreases in snow depth and snow cover duration are found in lowland and mid-elevation regions, while these trends are generally weaker at higher elevations and are partly not statistically significant. SNOWRAS provides a spatially consistent, high-resolution dataset for monitoring snow conditions and investigating long-term changes in Germany’s snow climate. Together with complementary information, such as snow water equivalent, the dataset can also support hydrological modeling, assessments of water availability, and analyses of hydrological risks.
The submitted manuscript presents on, the one hand, the methods used to create a new 1 km gridded climatological dataset of daily snow depth for Germany. On the other hand, shows the analysis of climatological mean values and temporal trends for some snow indicators derived from this long-term daily gridded dataset. The results focus on the differences between climatological norm periods and elevation dependent differences.
The topic is of high interest to readers of HESS and the new product will certainly be of broad use for many snow-related applications. I liked reading the manuscript, which has a clear structure and illustrative figures. The language is with a few exceptions easy to read. I suggest accepting the manuscript as soon as the following major points have been addressed.
Major points:
Minor points:
L23: Please rephrase the sentence “Snow-covered areas reflect...” because it currently contains a grammatical error and is difficult to understand.
L43: Frei et al. 2018 is about future snowfall and not snow depth! Newer studies which would be appropriate here are:
doi.org/10.1029/2018GL079799, doi.org/10.3389/feart.2023.1165861, doi.org/10.1002/joc.5902, doi.org/10.3354/cr00996
L50: Marty et al. 2025 and not 2017 is the appropriate reference.
L59: Klein et al. 2016 analyzed less than a dozen stations. More appropriate references would be:
doi.org/10.1002/joc.3653 or doi.org/10.1029/2008GL033998
L93: There is nothing on partnering networks in Kaspar et al. 2013. Please refer to Acknowledgements.
Fig 1: Please also add a figure which demonstrates the elevation distribution of the stations.
L 114: Unfortunately, in Kaspar et al. 2013, there is nothing specific on snow. It would be very valuable for other countries or researchers if more details could be provided in Supplements. For example, it is not at all clear if other measured variables like snowfall were also used to check the plausibility, if and how corrections were applied (manually or automatically) and if and how data gaps were filled?
L137: “snow-free” instead of “snow-poor”.
L150: Can you please provide some numbers for a typical station density.
L170: Does this mean stations above 800 m were not tested for spatial consistency?
L172: The excluded 11000 observations came from how many stations (in percentage)?
L192: Please specify if the selection of stations for each target grid cell was done daily, which would imply that the chosen stations can change from day to day.
L197: Please, first describe the general idea behind the approach and mention if the final interpolated values where rounded or not?
L198: horizontal location s = grid cell?
L200: Please, for Dres(s) make a reference to 2.4.3.
L211: The correlations between the “selected” stations…
L240: Missing data: Does this mean gaps were not interpolated at all?
L251: Snow cover period: The number of days between the first and last snow cover day, including intervening snow-free days.
L 260: Please provide some explanation, how seasons without any day with snow on the ground were considered for mean and maximum snow depth. As 0 cm or NAN?
L273: The remote sensing community would disagree at least for the indicator “number of snow days”, therefore I suggest showing a comparison even if it’s only for climatological means.
L360: Fig. S2: Why is there a thick zero line in the plot for Hohenpeissenberg? Does it only look like or has Breitnau after 2002/03 no differences anymore – the difference in 2006/07 seems to be an error as the measured series shows zero mean snow depth?
L366: Fig. S3: Please mention at least in Discussion why some grid cells, e.g. between Saarbrücken and Frankfurt, show positive differences?
Fig. S4: Same here for isolated small positive regions over all Germany.
Fig. S5: There is definitely information needed how the spatial domain of this mountains regions is defined and where they are located. Providing the median elevation of each region may also help to explain the different trends.
L399: This has already been shown by other studies (e.g. Klein et al., 2016 ; Matiu et al., 2021). Please reference in Discussion.
L405: “Cumulative increase in snow depth”: Please shortly explain why this new indicator and what it can be used for?
Fig. 8b: Please mention in Discussion, that this analysis shows no clear elevation dependence due the fact that only days with snow cover were considered.
Fig. 8d: “Seasonal number of snow cover days”: Do you mean “number of snow days” or “snow cover period”?
L473: Please explicitly note that these values for average snow depth cannot be compared with those from other studies, as days with a snow depth of 0 cm were not included.
L474: However, values for the number of snow days should be compared to similar studies like
doi.org/10.3354/cr00996 and doi.org/10.1029/2004GL020255.
L476: I guess the winter 2010/11 would be far less prominent, if mean snow depth would also consider snow depth =0 cm. If yes, it should be mentioned.
L491: As in L474, trend results should be compared with similar studies like doi.org/10.1029/2008GL033998, doi.org/10.5194/tc-15-1343-2021, doi.org/10.3929/ethz-b-000652501, and doi.org/10.5194/tc-19-4391-2025.
L493ff: It is not all clear if these sentences refer to snow depth or snow days?
L509: This is a repetition of L476.
L518: What is “seasonal snow amount”?
L525: It would be “Stewart, 2009” and more Germany related and newer references would be doi.org/10.3389/feart.2023.1165861
L565: omit “fully”
L569: omit “fully”