Impact of wind pattern and complex topography on snow microphysics during ICE-POP 2018 Author’s Responses to Reviewers

Please see below for our point-by-point response to all reviewers’ comments. Our responses are highlighted with a box filled in gray, followed by the corresponding changes to the manuscript. Changes that have been made are marked in :::: blue within the modified text represented by “...”, while the removals are highlighted with the deleted red text (for example, red). The figure and table numbers in our responses are based on the revised manuscript.

1.3. L173-176: 80% rain seems like a very large fraction of rain for an algorithm to eliminate rain. Perhaps 40% would be more appropriate?
The referee is correct. The threshold has been changed as suggested. The corresponding figures have been regenerated: Figures 5,6,7,8,9,10,11,and 12. This threshold change affected mostly to the number of data of GWU PARSIVEL due to a warmer environment (lower altitude). The changes of the results were minor and there have been no need to change any scientific logics to reach our conclusions. The only corrections needed was just to change a few numbers in the manuscript. Furthermore, we identified a few numerical typos in line 399-400 and amended them in this revised manuscript.
We have now clarified it in the text. The reason for the normalization by this value is to make CFADs from different sites to be comparable by dividing the same value.
Many thanks for the positive evaluation.
1.6. L404-408: The data suggests to me that the increased turbulence in the 3-3.5 km layer is not just causing increased aggregation, but also riming. Below 3 km in Fig 3, the velocity spectrum shifts to the right (faster fall velocities). This would be consistent with graupel now appearing in the spectrum of hydrometeors. Amended.
We agree with the reviewer that at a given height the warmer temperature is favorable condition for aggregation, however, in this study, the aggregated snow at the ground at GWU site was due to a warmer temperature of lower altitudes. Despite that, we have checked if the coastal sites were warmer than mountainous sites at a given height. The Fig. S1 shows the temperature profiles from upper-air soundings at three sites (DGW, BKC, and GWW) for the air-sea interaction event. Please see purple filled circles in Fig. 1 for the site location. The result suggests that the temperature of airmass did not decrease as the airmass moves to the mountainous region. The temperature at DGW was even about 1°C higher than those of coastal sites (Fig. S1a), indicating the aggregation at GWU is likely contributed by the warmer temperature of lower altitudes. 1.8. L448-450: The authors have shown very convincing evidence that both riming and aggregation increase from YPO to CPO, but I think this sentence does not quite support the argument. In my opinion, the way to make the argument for both riming and aggregation would be to say something like "the doppler velocity spectrum at CPO has a similar median to that at YPO, but the spectrum is much wider, suggesting an increased frequency of both slow-falling aggregates and faster-falling rimed particles (Fig 8)." The next sentence, mentioning Figure 9, seems great and does not need any change.
Revised. Thank you for the suggestion.
The ranges of the x-axis in the reflectivity CFADs have been reduced ( Figure 11). In addition, for the consistency, the x-axis ranges of reflectivity CFADs in Figures 4, 7, and 9 have been reduced. This was actually the reason why quantitative values have been provided in the sentence.
1.10. L494-495: I'm confused by this statement, because it appears to me in Fig. 11b that GWU has a higher frequency than YPO in the smallest Dm bin.
" A possible evidence of snowfall enhancement by seeder-feeder mechanism (Bergeron, 1965) may be found in the continuous increase in reflectivity of the primary peak (see normalized frequency > 2.5%). ::::::::::: Comparing mountainous regions (YPO and CPO) and coastal regions (GWU), we can see that the increase of reflectivity at the GWU site below 3 km ... " 1.12. L525: This paper (https://doi.org/10.1175/MWR2874.1) occurs in a somewhat comparable regime (sea-effect snowfall) and they document the tendency for increased riming over even relatively small peaks. It corroborates the results of the present study nicely.
According to the reviewer's suggestion, we added the Kusunoki et al. (2005) paper in the revised manuscript.
1.14. Fig. 1: if the size of the markers were reduced by 30%, the reader could see the terrain in the vicinity of the sites a bit more easily.
Corrected as suggested.
1.15. L445: All I see in the PDF I am reading is "m s −1 " with no number. This could be an error by the Copernicus website. Corrected.
" Note that the median wind speed of all cold low events was :::: 6.70 m s −1 at DGW located between YPO and CPO, while they were 2.60 and 2.30 m s −1 for air-sea interaction and warm low events, respectively." 1.16. L464: I can see "2.5 m s −1 ", but the other wind speed in this line says "m s −1 " with no number. This could be an error by the Copernicus website. Corrected.
" Both the frequency of Doppler velocity above ::: 1.5 m s −1 and primary frequency peak of Doppler velocity of an approximately 1.0 m s −1 rise in that layer (3.0 -3.5 km) and maintain their velocity as they fall." 1.17. Is it possible to list the mean liquid-precipitation-equivalent rate for each site in Figs 6, 8, and 10? It would be somewhat informative to know what these radar characteristics translate to as far as liquid rates.
Good point. Unfortunately, it was not able to get the liquid-equivalent precipitation rate measurements from collocated weighing gauges for the entire period of the studied events. For instance, the Pluvio weighing gauge at YPO was deployed from March 2017, meaning the observations of at least five events of the selected events (Table 1) in January and February are not available. Moreover, no weighing gauge was operated at GWU in 2016-2017 winter season. Instead, in response to this comment, we have calculated the mean liquid-precipitation-equivalent rate at 1.5 km height (for each site) from the Z e -S relationship obtained at the MRR wavelength (Souverijns et al. 2017). They are now listed in the captions of Figures 4, 7, 9, and 11.
The manuscript documents the microphysical characteristics of snow by analyzing the PARSIVEL and MRR data collected from ICE-POP 2017-2018. The snow events were classified by three different synoptic systems. The three type of synoptic systems are air-sea, cold low and warm low patterns. The results show distinct characteristics of snow. The aggregation process increases the size of the snow. The riming process has higher values of fall velocity of snow particle. Most of the conclusions are reasonable, but further detailed analysis is missing. However, the manuscript did a good job summarizing 20 snow events from ICE-POP 2017-2018.
We would like to thank the reviewer for careful reading of this manuscript and the positive review. While we are sorry to hear the indefinite comment "further detailed analysis is missing", we believe that we have proposed a new understanding on the impact of wind flow and topography on the microphysics in this region following the objective: "The purpose of this paper is to elucidate the microphysical characteristics of snow in the Gangwon region in both different heights and surface by using MRR and PARSIVEL datasets." Please note that the future study will cover more detailed view of microphysics and dynamics for each synoptic system by means of high-resolution radar-based products (e.g., 3D wind field) and polarimetric radar variables. A point-by-point response to the comment is given below. We hope the referee find our responses and the revised manuscript satisfactory.
Minor comments 2.1. Line 55: "showed that the mean precipitation amount increased by about 45% in the presence of both Kaema high and low pressure. " Any reference? Song et al. (2016) showed that the coexistence of the Kaema high and low pressure increases the mean precipitation amount by about 45%. From the pronoun "They" in the original manuscript, we believe the readers can easily recognize the reference from the manuscript: " Song et al. (2016) classified the synoptic environments.... They showed that the mean precipitation amount....." Thus, we have decided to keep the sentences in the current form without citing the same article additionally.
2.2. Line 419: "the size of the snow does not greatly enlarge." Why? Evidence?
It is obvious that the increase of the dimension during riming is not significant compared to aggregation process. While the size (or volume) does not increase much, the mass increases during riming, leading to increases of density and fall speed of snowflake, and the Doppler velocity. This statement is partially based on a conceptual model of the growth of snow by riming of Heymsfield et al. (1982) which has been adopted in the current microphysical parameterizations of riming (Morrison and Grabowski 2008;Jensen and Harrington, 2015;Morrison and Milbrandt, 2015). Based on the model, during riming, it is hypothesized that the rime is accreted to the interstices between the crystal branches while the maximum dimension is conserved until it becomes a spherical shape. Although a new parameterization was suggested recently because the increase of maximum dimension during riming was found from the simulation (Seifert et al. 2019), the increase should be limited to the graupel size of typically 2-5 mm (Heymsfield et al. 2018). The citation is added in the manuscript: " When riming occurs, Doppler velocity is increased because the mass of snow increases due to collection of supercooled liquid water, while the size of the snow does not greatly enlarge ::::::::::: (Heymsfield ::: et ::: al. :::::: 1982), resulting in nearly constant or smaller increase of reflectivity. " 2.3. Line 422: Where is the second peak of reflectivity?
Clarified. The second peak of reflectivity can be found at the lower reflectivity below 3 km height, while the primary peak appears at the reflectivity greater than 15 dBZ. " In the layer where the updraft is present, there is a secondary peak of reflectivity ::: (at :::::: lower :::::::::: reflectivity :::: less ::::: than ::: 12 :::::: dBZ) and enhanced spectral width in mountainous sites up to a height of 3 km. "