The manuscript entitled “Multi-year precipitation characteristics based on in-situ and remote

sensing observations at Ny-Ålesund, Svalbard”, present precipitation detection and measurements by several methods in the period 2017-2021 in Ny Alesund and show how atmospheric circulation (AR, cyclones, front) impacts the precipitations events. I think the paper addresses very important questions in an area of rapid change in the context of global warming. The paper is generally well written and the figures are clear and of good quality. However, I think the quality of the paper can be greatly improved by focusing on more specific questions, reorganizing a bit the different parts and expending on the discussions.

First, the introduction is very informative with a lot of cited literature. You introduce well the need of improving ground based observations. However, the connection between the previous studies, the need of new observations for precipitation in Ny Alesund, and the new measurements that you describe is not clear. What are exactly the measurements that were missing before the instruments were installed and how these new measurements can improve our knowledge on Ny Alesund precipitation? the questions that you mention in the introduction: “How much precipitation falls at Ny-Ålesund, and how is it related to the previously mentioned weather systems? What type of precipitation occurs? How often does it precipitate?” have been already addressed in previous studies. You need to point out what these new measurements are adding to the previous answers to these questions and what is new in your way of relating the precipitation events to weather systems. In summary, you need to be more specific on how these new measurements can help answer your research questions.

Second, I suggest reorganizing the manuscript, because in the present form, the flow of ideas is a bit confusing. I really like the part on explaining the weather patterns origins of precipitation events, but it comes too early in the manuscript. I suggest first introducing the results of the measurements (quantity of precipitation, types, frequency), and then the large-scale weather patterns. The advantage of this organization would be that you can use your previous results on separating snow and rain to study the role of weather patterns on rain and snow events. Summer rain, winter rain (rain and snow) and snow, have very different implications for other research studies and local communities, so it would be extremely interesting to use the pluvio and parsivel measurements to study the specific origins of rain and snow. This analysis would be an application of your method of separating rain and snow. Regarding the part on frequency and the use of MRR, I don’t really see the benefit on this study. It needs to be better explained or to be removed from the analysis.

Finally, your article lacks a discussion part where you talk about previous studies and show how the new measurements may improve the previous studies. Many articles are already cited in the introduction but they don’t serve your discussion in the current version of your manuscript. You have some discussion spread in the manuscript, but it would be good to write a specific discussion section and expend on them.

I think this manuscript have a great potential to get published after major revision, and is a very important step in improving our precipitation knowledge in Ny Alesund.

Here are some specific comments:

I suggest changing the title to “Multi-year precipitation characteristics based on in-situ and ground-based remote sensing observations at Ny-Ålesund, Svalbard”.

l.11 «Cyclones contributed 40% (21%) of the total precipitation amount if all

(separated) cyclone events are considered ». This sentence is not clear to me.

l.13-14: I am not sure this is needed. Of course the occurrence is lower when the resolution is 1 minute compared to daily. What is the resolution of micro rain radar?

l. 32-40 maybe separate what is observed in the past and what is modeled in the future, it is bit mixed up here. There is already literature on past increase of precipitation and rain in the Arctic that you can include.

l. 60-61 «melt days in winter in Svalbard and the associated precipitation sums have increased» not sure I understand what are precipitation sums and how it is associated to melt.

l. 61 «These rain-on-snow events». I would say «the rain on snow events» since you don’t talk about rain-on-snow events previously.

l.77-78: this sentence can be simplified. you say three time the same thing: «model data» «from numerical weather prediction models»and «climate simulations».

l.104: operated by?

l.140 What is the resolution of parsivel?

l.164: what is the horizontal resolution of the MRR?

l.188-189: you can just say june to september

l.231 it seems that the underestimation of your pluvio data compared to MET norway is larger in winter months. It would be interesting to also show the average difference per month.

l. 232-23 you can use the precipitation correction of the MET Norway time series by several methods here https://www.easydata.earth/#/public/metadata/a3d7b9e6-9626-4d43-bb83-623900eb1053. It would be good to include them in table 1 and figure 2. I suggest also to show a scatter plot with corrected MET norway vs corrected pluvio for daily values.

l.257: methodology of?

Table 2: AR-CY not AR-CA. The results are strongly dependent on the occurrence of these weather patterns. It would be good if overall occurrences appear somewhere.

l.259: you mean a monthly maxima?

Figure 4: it also seems that the combined AR-CY don’t happen frequently but are associated to a lot of precipitation. I think this needs to be discussed.

l.264-266: I think you should discuss the specific months after discussing the general behavior.

Figure 5: it would be better to write the bins in th x axis: [-0.5-0.5], [0.5-1.5], [0.5-1.5]… because I am not sure what bin corresponds to each point here. For example, does 0 corresponds to [-0.5-0.5]?

l.276: I question the choice of including hail in solid precipitation as snow and hail are driven by very different processes and may not occur in the same season.

l. 280 I think MET Norway also gives the type of precipitation: snow, liquid, mixed. It may be good to compare with this. And also to compare with snow-rain separation done by Champagne et al. 2024 using a threshold of 1°c (dataset above). I think your study as the potential to determine more precisely a threshold for solid vs liquid precipitations that can validate previous datasets or be used in future studies. It is great and can be emphasize in your manuscript.

l.283 «this makes sense» is a bit familiar. Temperature is mostly used because there is no alternative and because temperature and precipitations are often long records measured together. In my opinion, what is very relevant your study is that the temperature threshold can be validated.

l.285: When you say 2°c you mean the 1.5c-2.5°c bin? Since you use a 1minute resolution, with a lot of data available, I would suggest using more bins, at least for critical temperature between 0 and 3°c (0.1°c wide bins?). You can then arrive with a suggestion of thresholds (or a ranged threshold) separating solid, liquid and mixed precipitation.

l.293-294: Not sure I understand when you say «during periods when solid precipitation has been detected by Parsivel as well (not shown)». parsivel have shown mixed precipitation during these events then?

l.294: what the size of particle infers for the type of precipitation? Expend on that.

l.314-315: I think the thresholds for mixed precipitation can be refined. Looking at figure 5, more than 90 % of precipitation is only solid below 0°c and only liquid above 3°c. Keep in mind that for most studies mixed precipitation is a useless variable. For most fields, we must know if snow is falling and sticking to the ground or if the precipitation are liquid (or almost liquid at the surface). The Parsivel is a very powerful tool that should be used at its best capacity to refine the temperature threshold between liquid and solid precipitation.

Figure 6: why for mixed phase the fraction of precipitation occurrence is way lower than the fraction of precipitation amount?

You could also look at the dataset from Champagne et al. 2024 https://www.easydata.earth/#/public/metadata/a3d7b9e6-9626-4d43-bb83-623900eb1053 and validate if a simple way of separating snow and rain (temperature threshold) can be done. your study can be very useful to validate such dataset.

l.323: «Since the mixed-phase precipitation type, as detected by Parsivel, only seldom occurs» awkward wording.

l.326: I think you can reduce the number of unknown cases by narrowing down the threshold [-2-4°c].

l. 333 i don’t see any analysis on ROS events.

l.345: >0mm or 1mm ?

l.348-349 «To understand which rain or snowfall rate a value of -5 or -10 dBZ actually represents, Ze − R and Ze − S relationships must be applied » Not clear. Reformulate.

l.370: Even though these frequencies serve different purposes, I think you still need to discuss a bit if these frequencies are plausible or not. The MRR frequency seems very high, but pluvio rather low. what causes the overestimation in MRR and Parsival? Also expend on the purpose of these different measurements (maybe in a separate discussion part).

Table 4: I assume these are data from pluvio? Are these raw or corrected data? It would be nice also to see the data from corrected Norway MET data.

l.375-386: What is the main benefit of this MRR for precipitation detection? I don’t really see it here. You are not really describing the vertical variability of the results and how it can be used. This needs to be discuss more.

l.393 not clear. You mean you need to have at least 30 minutes with a Ze < -10dBZ (no precipitation) between two events?

l.394: an event interrupted by for example 12h without precipitation is also likely from the same weather system. So the 30 minutes needs to be better justified.

l.395 discussed later in another paper you mean? I don’t see this discussion.

l.396-397 «Here, we extended the considered period by ±10 min due to the spatial distance of the Pluvio and MRR measurements» you mean that rain may take 10 minutes to reach one site to another? This needs to be better explained.

l.403-406 I don’t really understand the goal of figure 9.

l 410: «maximally contribute to the total precipitation amount with 11%» not clear

l.407-421: The results are interesting but I question what is the end goal of showing all these results. The longer and stronger events of course are more rare but give more precipitation amounts. What is new about it? What are the implications of these results? More generally I have hard time to see the benefit of using MRR here.

l.424 A follow up study would be very interesting as your article leads to lot of new questions. I think it would be very interesting to separate the type of extreme events (summer rain, snow, rain on snow) and to focus on ROS events. It can be done here and would show a good application of pluvio-parsivel.

l.464-472 I would first summarize your main results before introducing what may be done in the future

l.473 replace small by low

l.483-485 I think from your results you can already give approximate temperature threshold for rain, snow and mixed precipitation at Ny Alesund. This is extremely useful for future studies and this needs to be higlighted in your article.

I suggest adding a discussion section and the conclusion needs also to be rewritten with more emphasis on very important results. In the current form the conclusion repeats a lot of previous results but it is hard to see what are the main conclusions.

General Comments

The study investigates precipitation characteristics at Ny-Ålesund, Svalbard, based on ground-based observations from three instruments (Pluvio, Parsivel and MRR), and relates precipitation to weather systems such as atmospheric rivers, cyclones, and fronts. It also investigates extreme precipitation events and describes one precipitation event in more detail. 

The manuscript is generally well written and logically structured, and presents a comprehensive dataset. The data analysis is mainly sufficient, figures are of good quality, and conclusions are supported by the data presented. However, there are some areas which need major improvements, particularly the clarity of the study’s motivation, the description of methods, and discussion of implications, see my specific comments. The text should be substantially clarified and improved. With improvements, this study will make a valuable contribution to understanding precipitation in the Arctic.

Specific comments

Introduction:

1. The introduction gives a strong rationale but it is quite general. It should better connect the objectives of this study to gaps in the literature. 

Methods:

2. The correction function for precipitation by Wolff et al (2015) is based on measurements made at Haukeliseter, Norway (at above 1000m altitude). How well does this correction function apply to the conditions in Ny-Ålesund? How did you select this correction function? There are plenty of other correction functions in the literature (see for example Kochendorfer et al. 2017, or Køtzow et al. 2020). Discuss the assumptions or limitations of the selected function and motivate the selection briefly. 

Kochendorfer, J. et al: Analysis of single-Alter-shielded and unshielded measurements of mixed and solid precipitation from WMO-SPICE, Hydrol. Earth Syst. Sci., 21, 3525–3542, https://doi.org/10.5194/hess-21-3525-2017, 2017.

Køltzow, M. et al: Verification of Solid Precipitation Forecasts from Numerical Weather Prediction Models in Norway. Wea. Forecasting, 35, 2279–2292, https://doi.org/10.1175/WAF-D-20-0060.1. 2020

3. Clarify which instruments are used for specific characteristics (precipitation amount, frequency, and type). For instance, in Section 2.3, it is currently unclear why MRR data is used or how it complements those from Pluvio and Parsivel.

4. The abbreviations O-AR, O-CY, etc., are introduced in Section 2.4 but not clearly explained. I suggest briefly clarifying these terms when they are first mentioned (O-AR refers to 'only atmospheric rivers'). Additionally, as the use of ERA5 data and the detection of weather events are central to the analysis and conclusions, you should consider creating a dedicated subsection to explain the detection algorithms and their relevance to the study. This would enhance clarity.

5. Why was Geonor not used in this study? Wouldn’t it provide an additional comparison or insights into precipitation measurement uncertainties?

Section 3

6. Small precipitation amounts are the focus of Figure 3 and the paragraph starting on line 240. The importance of this analysis is not clearly motivated. Please explain why studying these small precipitation events is relevant.

7. Lines 290–302 discuss mixed-phase precipitation but do not address sub-zero liquid precipitation, which also occurs. Is this considered in your analysis?

8. Could additional data sources validate precipitation types from Parsivel? For instance, the video data you mention or manual observations (SYNOP/METAR)?

9. Line 346: How was the height of 120m selected? Can precipitation detected at this height reliably represent surface precipitation? What uncertainties are associated with this approach? Discuss ,and modify the methods section accordingly.

Section 4

10. Line 401: "for which Ze values precipitation is actually detected on the ground" is a critical question that applies to previous sections as well (e.g., precipitation frequency analysis). Please address this earlier in the text.

11. Lines 401–406: Results are described but not explicitly interpreted. Connect these findings to the question you are trying to answer. For example, explicitly address how these Ze values relate to surface precipitation detection.

12. Figure 10: I am not sure about the journal recommendations for figures, but for me, it would be easier if a legend was added directly in the figure, not just the caption. This applies to other figures as well, but it is especially important here because of the number of lines in the plot.

13. Captions in general: Indicate which data sources the figures are based on. While this becomes clear when reading the text, captions alone should provide this information.

14. The analysis of extreme precipitation cases is valuable, but the manuscript refers too much to future studies. For example, on line 441, avoid starting with a reference to a follow-up work. Focus on the results gained from these cases now, such as the role of ARs.

Conclusions:

15. Clearly state the main conclusions. Right now, they are mixed with comments about limitations and future work, which makes it hard to focus on the key messages. I suggest putting limitations and future work into separate paragraphs.

16. Line 473: The statement "Large-scale weather events like ARs and cyclones are common features at Ny-Ålesund" appears to contradict the statement on lines 473–474 about ARs occurring only 8% of the time. Please clarify.

17. Line 486: The reference to "In many studies" is vague. Specify which studies base precipitation type solely on temperature thresholds.

18. Discuss how your findings contribute to current knowledge about Arctic precipitation, as introduced earlier in the manuscript. How do they advance our understanding of precipitation processes in the Arctic climate? I would like to see more discussion on what these results mean and what the community can learn from this study.

19. Specifically, you mention model uncertainties in the introduction. Could your observations help improve understanding and even improving models? What are the potential applications for the modeling community?

Minor issues 

Line 27: “the Svalbard archipelago … reveals the highest temperature increase” sounds strange. → The Svalbard archipelago exhibits, or has experienced. 

Figure 2 caption: Clarify that the hatched areas denote months with significant data gaps.

Line 475: Remove the extra ) after Ny-Ålesund.