the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 14 Aug 2026
Local Controls on the Development of Percolation Features After a Rain-on-Snow Event in a Southern Taiga Snowpack
Abstract. Rain-on-snow (ROS) events can rapidly introduce liquid water into cold snowpacks. However, direct field observations of preferential-flow structure development under such conditions remain limited. In this study, we document the formation of percolation columns and linear flow pathways after a ROS event in a cold, predominantly fine-grained snowpack in the southern taiga. We combine meteorological observations, snow-pit measurements, stratigraphic descriptions, density and snow-water-equivalent (SWE) measurements, and low-cost photogrammetric mapping to examine how local surface conditions influenced the spatial organization of preferential flow.
The same ROS event produced contrasting flow structures over short distances. In the floodplain meadow, percolation features developed mainly in inter-tussock depressions and formed relatively large funnels with the highest areal coverage. In the more homogeneous field site, features were more numerous but smaller, while in the birch forest they were sparse and clustered, mainly in canopy gaps. On the slope (~25°), liquid water was routed into elongated, downslope flow pathways rather than discrete vertical columns. At the meadow and field sites, snow within percolation columns was denser and stored more SWE than the surrounding snow occupying the same volume, indicating localized water accumulation and refreezing.
These results show that rain-induced infiltration in cold snowpacks can be reorganized at sub-meter scales by microtopography, vegetation, slope, and snow stratigraphy. Percolation columns and associated ice structures therefore represent important small-scale heterogeneities that can affect meltwater storage, refreezing, drainage pathways, causing challenges in snowpack modelling, and the interpretation of microwave remote-sensing signals.
- Preprint
(1225 KB) - Metadata XML
-
Supplement
(4778 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2026-3873', Anonymous Referee #1, 13 Sep 2026
-
AC1: 'Reply on RC1', Anton Komarov, 23 Sep 2026
Response to the major comment
We thank the referee for this important comment and for recognizing the effort involved in obtaining detailed observations during a naturally occurring ROS event.
We agree that observations from a single ROS event cannot determine how the snowpack response would change under substantially larger rainfall amounts, different rainfall intensities, deeper snowpacks, or other initial conditions. For example, an event delivering 10, 20 or 30 mm of rain to a substantially thinner/deeper snowpack could produce different infiltration, storage, snowmelt and drainage behaviour. We have therefore revised the manuscript to state this limitation more explicitly and to avoid implying that the quantitative characteristics observed during this event can be generalized to all ROS conditions.
The main objective of this study was not to establish a generalized relationship between ROS magnitude and snowpack response, but to examine how the same regional ROS event produced contrasting percolation patterns under different local conditions. Across four nearby sites, preferential-flow structures varied markedly with microtopography, vegetation, slope, and snow stratigraphy, ranging from large funnels in the tussock meadow to numerous smaller features in the field, sparse clustered features in the forest, and elongated downslope pathways on the slope. The present study therefore documents a relatively small ROS event over a moderately thick seasonal snowpack, while the range of conditions over which these relationships remain valid is still unknown. We have revised the manuscript to make this scope more explicit and to emphasize the need for further observations, controlled experiments, and modelling across a broader range of ROS conditions.
A second contribution of the study is methodological. We combine conventional snow-pit, density, SWE, and stratigraphic measurements with low-cost smartphone-based mapping using Scaniverse. This approach allowed us to characterize the size, density, spacing, and spatial organization of preferential-flow features at the plot scale, providing information that is difficult to obtain from conventional snow-pit observations alone. We also evaluated the reliability of the derived spatial measurements against independent field measurements. The Abstract and Introduction have been revised to emphasize this methodological contribution more clearly.
We agree that additional observations, controlled rainfall experiments, and numerical modelling would be valuable for determining how these patterns change with rainfall amount and intensity, snow depth, initial snow temperature, stratigraphy, and other boundary conditions. Addressing this broader range of conditions, however, would require a separate experimental and modelling framework beyond the scope of the present observational case study.
We believe that these revisions clarify that the contribution of the study is to document spatial heterogeneity in the snowpack response to a natural ROS event, identify potential local controls on that heterogeneity, and demonstrate a practical low-cost approach for mapping these features, rather than to derive a universal response applicable to all ROS events.
Response to Specific comments
Comment 1: The manuscript lacks a map of the study area and photographs of the surface conditions at the field observation sites.
We agree that this information is important for understanding the differences among the four study sites. We have added a study-area map showing the location of the field site. Photographs illustrating the characteristic surface, vegetation, and microtopographic conditions at the sites are already presented in Appendix 1.
Comment 2: The font size in the hand-drawn diagram is too small, and the font styles are inconsistent. Please improve the visual presentation of the diagram.
We have revised the figure. The labels have been enlarged, font styles standardized to improve readability and consistency.
Comment 3: The discussion should be expanded and should include an uncertainty analysis and aspects not covered by the manuscript.
We agree with this comment. We have substantially expanded the Discussion to distinguish more clearly between observed results and process interpretations and to address the major sources of uncertainty. In particular, we now discuss the limitations associated with investigating a single ROS event; the relatively modest rainfall amount; the limited range of snow depths and initial snow conditions represented by this event; differences in pre-event snow stratigraphy among the sites; the absence of direct measurements of liquid-water flux during infiltration; uncertainty related to the temporal separation between the ROS event and post-event observations; manual identification of percolation features; and the spatial accuracy and scale limitations of the smartphone-based mapping approach.
We have also expanded the outlook to discuss how controlled rainfall experiments and numerical modelling could systematically investigate the effects of rainfall amount and intensity, snow depth, initial snow temperature, snow microstructure and stratigraphy, and landscape characteristics. We emphasize that such work is required before the quantitative relationships observed during the present event can be generalized to other ROS conditions.
Comment 4: Please standardize the literature format.
Thank you. The reference list and in-text citations have been revised and standardized according to the HESS formatting requirements.
Citation: https://doi.org/10.5194/egusphere-2026-3873-AC1
-
AC1: 'Reply on RC1', Anton Komarov, 23 Sep 2026
-
RC2: 'Comment on egusphere-2026-3873', Benjamin Bouchard, 23 Sep 2026
In the manuscript by Komarov and Stroeve entitled ‘Local Controls on the Development of Percolation Features After a Rain-on-Snow Event in a Southern Taiga Snowpack’, the authors analyze the snowpack characteristics at for locations before and after a rain-on-snow event that occurred in the southern taiga of Russia in December 2024, in particular the formation of percolation channels.
The study is original as it reveals macroscale details that are often overlooked by the ssnow science community. Although the study case presented here is interesting, the manuscript is underachieving and must undergo major changes before being considered for publication in HESS. My main concerns regarding this manuscript are summarized as follow:
- It must be clear that this work is a case study and not a general conceptualization of the effects of ROS on snowpack features.
- The methods used in this study must be better supported and detailed
- Some analyses are lacking, in particular the assessment of uncertainty.
- Limitations and perspectives of this work (Discussion) are currently too weak for publication and must be strengthened.
Here are numerous specific comments (minor and major) that should be addressed before submitting the manuscript for another revision.
Specific comments:
- 32: Please correct the reference in the text: Pomeroy and Brun, 2001.
- 41 to 44: I don’t see the opposition between these two statements. I suggest rephrasing to emphasize that capillary barriers can form in the presence of a fine-to-coarse layer superposition (see Avanzi et al. 2026 (https://tc.copernicus.org/articles/10/2013/2016/))
- 62 to 64: In my view, this paragraph essentially addresses vertical ice features in the snowpack. However, this sentence mentions exclusively an effect of the horizontal structures such ice crusts of internal ice layer on wildlife access to resources. I recommend you reorganize this part of your introduction to present sequentially the effects of 1) horizontal structure and 2) vertical percolation features.
- 75 to 81: I suggest to clearly state the objective of your work at the end of the introduction: ‘Overall, the objective of this study is to …’.
- 82 to 85: I like this idea of explicitly describing very close expressions in the introduction, but I think that it would improve the readability if these expressions were defined in the running text after the first mention.
- Not that I want or ask you to cite my work, but I think that your manuscript is closely related to Bouchard et al. 2024 (https://doi.org/10.5194/tc-18-2783-2024), in particular, in section 4.1.2 where snowpack observations following ROS are thoroughly documented and supported with photos. Based upon the analysis and the hypotheses exposed by Bouchard et al. (2024), how does your study contribute to deepening the understanding of the impact of ROS in boreal forests? I think that this would reinforce the introduction and the first part of the Discussion. Other studies relevant to your work include Teich et al 2019 (https://doi.org/10.1016/j.foreco.2019.01.052) and Bouchard et al 2022 (https://doi.org/10.1002/hyp.14681). Perhaps you can have a look at these.
- 86: I recommend sorting this section (Materials and methods) into two subsections: 1) study site and 2) Experimental protocol. Moreover, this section should include a map of the site. Perhaps, you can move figure S1 to your study site subsection and complement it with an overview of the site location relative to the three weather stations and the town of Khotkovo with a relevant background map.
- 87: Please mention the coordinates of your site (l. 101) here.
- 89 to 91: Please be more specific here. What is the mean winter temperature and precipitation here?
- 91 to 93: How variable are snowpack properties from year to year? Some references are indicated but it would be necessary to mention the information explicitly here.
- 96: at what height are the met data measured?
- 103: Is the site surrounded by topographic features (e.g. hills, mountains) that would alter wind or radiation?
- 105: what height are the surrounding trees?
- 111: Please indicate explicitly the dates of snow pit measurements
- 115. Were snow density and temperatures measured as profiles (at what vertical precision?) or based on the identified snow layers? Please specify.
- 116 to 117. This information should appear later (l. 130) where you describe the related protocol.
- 125 to 128: this should be included in the running text.
- 130 to 135. The description of your protocol here is slightly elusive. I would expect more precision on the morphological characteristics that you aim to quantify. Is it the length of the percolation fingers, their diameter, the tortuosity? Also, as you mention, this approach includes important limitations (l. 134). The uncertainty on the estimation of the density, SWE and contribution to the mass balance should be quantified.
- 136: a reference or a link to the software should be added to your text.
- 136 to 143: Please detail more thoroughly what Scaniverse does and how it is used to identify the percolation funnels. A reference would be helpful here. Does Scaniverse use a objective threshold to determine or is it set by the user? Also, be more specific about the analyses that you perform with the ‘Python libraries’. Maybe the link to Zenodo should be already available.
- 152: As a scalar, wind speed is bounded to 0 m s-1 by definition. Saying that it fluctuated between 0 and 5 m s-1 sounds a bit clumsy. Maybe say that wind speed reached 5 m s-1.
- 155 this sentence is not useful. Please remove.
- 150 (Weather conditions): This whole section includes results that do not belong here as are described the effects of weather on snow stratigraphy (l. 155 to 162; l. 163 to 165; 167- 170). These results should be moved to the following section, and this section 3.1 should be thoroughly reorganized to include only results related to weather conditions: temperature, wind speed, precipitation and snow depth.
- 171 (Figure 1): Where does snow depth measurement shown in Figure 1 come from? Was it measured at one of the weather stations or is the snow depth time series retrieved from your field measurement; I am not sure if this was mentioned. Also, can you explain why the initial snowfall caused the formation of a compact basal layer (l. 157) but Fig 1c shows that snow completely melts out on December 2?
- 175 (Figure 2): I think that this figure should be removed as it repeats what is shown in Figure 1.
- 179 to 180: This was already mentioned in Methods. Please remove.
- 178 (Section 3.2): This section essentially presents how the snow stratigraphy evolves until the ROS of 21 December. It gets a bit hard to follow since all 4 sites are presented at once. Maybe you should specify at which location were measured the lower and upper bounds of snow density ranges that you indicate.
- 183 to 184: Is this what you show on Figure 1. Please specify.
- 198 (Figure 3); Here you only show snow profiles after the ROS despite describing mostly conditions before in the text. It would be insightful to have the profiles (grain type, density, SWE, temp) before and after. Moreover, The density profile should be presented as discretized values as each point is representative of one layer. Then I don’t think it would be necessary to present the profile of SWE, but rather indicate the bulk SWE of the snowpack (before and after ROS).
- 203 to 204: Please review the syntax of this sentence
- 207 to 210: This belongs to methods. Please remove.
- 212 (Figure 4): Were the KDE maps normalized. I find it curious that it reaches 1.
- 226 to227. The standard deviation of the nearest-neighbor distance would be needed here to support this statement.
- 228 to 230: Can you elaborate on the Ripley’s L-Function and the CSR? Maybe it is just me, but this approach does not seem to be meanstream. I think many readers (like me) may get lost here.
- 232 to 234: I feel like this belongs more to the discussion than the results
- 241: Very interesting!
- 255: Please elaborate on that and specify what you mean by ‘directly influenced by tree canopies’
- 272: It would be interesting to show a photo with the analysis from Scaniverse of these elongated flow pathways similar that shown on the left side of Figure 4.
- 275: What is the aspect of the slope?
- 280: Study from Eiriksson et al 2013 (https://doi.org/10.1002/hyp.9666) is of interest here.
- 282 (Sectoin 3.4). In addition to the schematic diagram (Figure 6), It would be relevant to add photographs mentioned earlier (l. 131) in this section.
- 297: Figure 6
- Panel c (forest site) should be removed as this was not supported by field evidence (at least not that is mentioned in the text).
- Indicate clearly on top of each frame which is the meadow and which is the field.
- If you have any photos of the snow pit wall showing the percolation columns (such as Fig. 5 from Bouchard et al. 2024), add it here.
- Can you explain why in Figure 6a, you have dry FC on the same layer as wet FC next to your percolation column? Please also correct FC crystals (the C already stands for crystals). Also specify what do you mean by wet FC, as FC is typically specific to dry snow.
- 309: Can you explain how macropores were formed in the snowpack before it got wet?
- 314: Can you confirm that you have observed more than only one column at each site, and perhaps rephrase? As I read this, it sounds that only one column at the meadow and fields sites was identified and this wouldn’t be enough to draw a generalized conclusion.
- 321: I don’t understand here. Did you assess percolation columns also at the forest site or not? It is confusing.
- 328: have you observed merging of preferential flow pathways at transition layers like Bouchard et al 2024 did?
- 344 (Figure 7): How I interpret this figure is that after some lateral distance water penetrates the snowpack vertically under homogeneous flow rather than preferential flow as is the case for the meadow and field site. Am I correct? If so, this should be somehow highlighted as it would be, in my opinion, an important finding of your work. Also, have you observed from a snow wall dug in the direction of the slope these ladder-like features of percolation or is this what you infer from your three snow pit observations? This should be specified.
- 346 to 362: This should appear before Figure 7 and the corresponding paragraph.
- 366: If liquid water refreezes on contact with snow surface is either because it was freezing rain or because the snow surface was below zero. Looking to Figure 2, We see that the temperature rises above zero before precipitation is recorded within the identified ROS period. Can you elaborate on that and address the role of pre-ROS snowpack conditions on water transport (vertical and lateral). In other word, what one can expect from ROS in the specific locations that you have presented if we were later in the season or in spring, when the snowpack is at 0 C?
- 386: I highly recommend you expand the discussion on limitations as these are numerous in your study. This starts with the fact that your conclusions rely on a single event at only one site. How can this be transposed to other locations under different ROS conditions? Also, your assessment of percolation column characteristics does not rely on any proven protocol and no quantification of uncertainty is provided. This lacks robustness and should be corrected.
- 406 to 410: The low-cost attribute of the photogrammetry approach is emphasized as a strength of this work (also in the abstract) but in fact, the Scaniverse software used in this study is not supported by any reference how can we know that this method was proven and validated? This must be address for an eventual publication.
- 405 (4.3 Broader use and future potential: The only point that is discussed here is the use of low-cost phone-based photogrammetry. However, I don’t think that this is the most important takeaway of your work. In fact, I would expect to discuss how your findings, i.e. the differences observed between locations, can be used in snow, hydrological and land surface models. How can this be integrated into a multi-layer snow models and how to upscale those sub-grid processes at larger modeling scale? I invite the authors to refer to the work of Mazzotti et al (2024) for some ideas (https://doi.org/10.5194/tc-18-4607-2024)
- 451 (Bibliography). Many items in the bibliography are not cited in the text (I have counted 3). Please make sure that the list of references matches the entries the manuscript.
- Except for Fig. S1, none of the figures in the supplement are indicated in the manuscript.
Citation: https://doi.org/10.5194/egusphere-2026-3873-RC2 -
AC2: 'Reply on RC2', Anton Komarov, 25 Sep 2026
We thank the referee for the detailed and constructive review and for recognizing the originality of the observations and the objective of the study. We agree that the manuscript should be framed more clearly as an event-scale case study rather than as a general conceptualization of snowpack response to ROS.
Accordingly, we will revise the Abstract, Introduction, and Discussion to emphasize the limits of generalization from a single natural ROS event. We will also reorganize the Materials and methods section to clarify the study sites, observations, measurements, and the distinction between directly measured characteristics and post-event interpretations.
We will strengthen the treatment of uncertainty by quantifying measurement variability where possible and by discussing the limitations associated with the single-event design, site differences, manual feature delineation, and destructive measurements of individual percolation columns.
The Discussion will be reorganized to compare the observations more explicitly with previous ROS studies, distinguish observations from process interpretations, and discuss implications for multilayer snow, hydrological, and land-surface modelling. Smartphone-based mapping will be retained as a secondary methodological contribution.
We will also address each specific comment individually and revise the corresponding text, figures, and references as appropriate. These changes will improve the methodological transparency, treatment of uncertainty, and positioning of the manuscript as a detailed field case study of spatially heterogeneous snowpack response to a natural ROS event.
Citation: https://doi.org/10.5194/egusphere-2026-3873-AC2 -
AC3: 'Reply on RC2', Anton Komarov, 04 Oct 2026
Dear Reviewer,
Thank you once again for your detailed and specific comments. We are continuing to revise the manuscript. Addressing each individual comment, including revisions to the figures, has taken somewhat longer than expected. However, we plan to finalize our response within the next few weeks.
Status: closed
-
RC1: 'Comment on egusphere-2026-3873', Anonymous Referee #1, 13 Sep 2026
Major comments:
This manuscript analyzes the characteristics of snowpack and the distribution of internal liquid water under different surface conditions in the southern taiga based on a single ROS event, which required the authors to invest significant human and material resources. The work deserves recognition. However, judging by the manuscript’s title, its primary aim is to examine the impact of ROS events on snowpack characteristics and the distribution of internal liquid water under different surface conditions. It is undeniable that a single ROS event can provide some insights. However, as shown in Figure 1, this ROS event was a small-scale one, with daily rainfall of approximately 5 mm. If precipitation were to increase to 10, 20, or even 30 mm, and snow depth reached 50 cm or greater, would this phenomenon still occur? Of course, this cannot be studied through natural ROS events—at least not in the short term—because the occurrence of ROS events is random and unpredictable. However, it can be determined through simulated rainfall experiments and model simulations. We recommend that the authors expand upon this manuscript to make its findings more widely applicable. Unfortunately, the current manuscript does not meet the journal’s quality requirements.
Specific comments:
- The manuscript lacks a map of the study area and photographs of the surface conditions at the field observation sites.
- The font size in the hand-drawn diagram is too small, and the font styles are inconsistent. Please improve the visual presentation of the diagram.
- In the discussion section of the manuscript, it is necessary not only to explain the causes of the observed phenomena, but also to expand upon aspects not covered in the manuscript and to conduct an analysis of uncertainty. As observational data are limited, and even models cannot account for all possible scenarios, it is essential to summarize the existing data and to provide an in-depth analysis and outlook on matters not addressed in the manuscript.
- Please standardize the literature format.
Citation: https://doi.org/10.5194/egusphere-2026-3873-RC1 -
AC1: 'Reply on RC1', Anton Komarov, 23 Sep 2026
Response to the major comment
We thank the referee for this important comment and for recognizing the effort involved in obtaining detailed observations during a naturally occurring ROS event.
We agree that observations from a single ROS event cannot determine how the snowpack response would change under substantially larger rainfall amounts, different rainfall intensities, deeper snowpacks, or other initial conditions. For example, an event delivering 10, 20 or 30 mm of rain to a substantially thinner/deeper snowpack could produce different infiltration, storage, snowmelt and drainage behaviour. We have therefore revised the manuscript to state this limitation more explicitly and to avoid implying that the quantitative characteristics observed during this event can be generalized to all ROS conditions.
The main objective of this study was not to establish a generalized relationship between ROS magnitude and snowpack response, but to examine how the same regional ROS event produced contrasting percolation patterns under different local conditions. Across four nearby sites, preferential-flow structures varied markedly with microtopography, vegetation, slope, and snow stratigraphy, ranging from large funnels in the tussock meadow to numerous smaller features in the field, sparse clustered features in the forest, and elongated downslope pathways on the slope. The present study therefore documents a relatively small ROS event over a moderately thick seasonal snowpack, while the range of conditions over which these relationships remain valid is still unknown. We have revised the manuscript to make this scope more explicit and to emphasize the need for further observations, controlled experiments, and modelling across a broader range of ROS conditions.
A second contribution of the study is methodological. We combine conventional snow-pit, density, SWE, and stratigraphic measurements with low-cost smartphone-based mapping using Scaniverse. This approach allowed us to characterize the size, density, spacing, and spatial organization of preferential-flow features at the plot scale, providing information that is difficult to obtain from conventional snow-pit observations alone. We also evaluated the reliability of the derived spatial measurements against independent field measurements. The Abstract and Introduction have been revised to emphasize this methodological contribution more clearly.
We agree that additional observations, controlled rainfall experiments, and numerical modelling would be valuable for determining how these patterns change with rainfall amount and intensity, snow depth, initial snow temperature, stratigraphy, and other boundary conditions. Addressing this broader range of conditions, however, would require a separate experimental and modelling framework beyond the scope of the present observational case study.
We believe that these revisions clarify that the contribution of the study is to document spatial heterogeneity in the snowpack response to a natural ROS event, identify potential local controls on that heterogeneity, and demonstrate a practical low-cost approach for mapping these features, rather than to derive a universal response applicable to all ROS events.
Response to Specific comments
Comment 1: The manuscript lacks a map of the study area and photographs of the surface conditions at the field observation sites.
We agree that this information is important for understanding the differences among the four study sites. We have added a study-area map showing the location of the field site. Photographs illustrating the characteristic surface, vegetation, and microtopographic conditions at the sites are already presented in Appendix 1.
Comment 2: The font size in the hand-drawn diagram is too small, and the font styles are inconsistent. Please improve the visual presentation of the diagram.
We have revised the figure. The labels have been enlarged, font styles standardized to improve readability and consistency.
Comment 3: The discussion should be expanded and should include an uncertainty analysis and aspects not covered by the manuscript.
We agree with this comment. We have substantially expanded the Discussion to distinguish more clearly between observed results and process interpretations and to address the major sources of uncertainty. In particular, we now discuss the limitations associated with investigating a single ROS event; the relatively modest rainfall amount; the limited range of snow depths and initial snow conditions represented by this event; differences in pre-event snow stratigraphy among the sites; the absence of direct measurements of liquid-water flux during infiltration; uncertainty related to the temporal separation between the ROS event and post-event observations; manual identification of percolation features; and the spatial accuracy and scale limitations of the smartphone-based mapping approach.
We have also expanded the outlook to discuss how controlled rainfall experiments and numerical modelling could systematically investigate the effects of rainfall amount and intensity, snow depth, initial snow temperature, snow microstructure and stratigraphy, and landscape characteristics. We emphasize that such work is required before the quantitative relationships observed during the present event can be generalized to other ROS conditions.
Comment 4: Please standardize the literature format.
Thank you. The reference list and in-text citations have been revised and standardized according to the HESS formatting requirements.
Citation: https://doi.org/10.5194/egusphere-2026-3873-AC1
-
RC2: 'Comment on egusphere-2026-3873', Benjamin Bouchard, 23 Sep 2026
In the manuscript by Komarov and Stroeve entitled ‘Local Controls on the Development of Percolation Features After a Rain-on-Snow Event in a Southern Taiga Snowpack’, the authors analyze the snowpack characteristics at for locations before and after a rain-on-snow event that occurred in the southern taiga of Russia in December 2024, in particular the formation of percolation channels.
The study is original as it reveals macroscale details that are often overlooked by the ssnow science community. Although the study case presented here is interesting, the manuscript is underachieving and must undergo major changes before being considered for publication in HESS. My main concerns regarding this manuscript are summarized as follow:
- It must be clear that this work is a case study and not a general conceptualization of the effects of ROS on snowpack features.
- The methods used in this study must be better supported and detailed
- Some analyses are lacking, in particular the assessment of uncertainty.
- Limitations and perspectives of this work (Discussion) are currently too weak for publication and must be strengthened.
Here are numerous specific comments (minor and major) that should be addressed before submitting the manuscript for another revision.
Specific comments:
- 32: Please correct the reference in the text: Pomeroy and Brun, 2001.
- 41 to 44: I don’t see the opposition between these two statements. I suggest rephrasing to emphasize that capillary barriers can form in the presence of a fine-to-coarse layer superposition (see Avanzi et al. 2026 (https://tc.copernicus.org/articles/10/2013/2016/))
- 62 to 64: In my view, this paragraph essentially addresses vertical ice features in the snowpack. However, this sentence mentions exclusively an effect of the horizontal structures such ice crusts of internal ice layer on wildlife access to resources. I recommend you reorganize this part of your introduction to present sequentially the effects of 1) horizontal structure and 2) vertical percolation features.
- 75 to 81: I suggest to clearly state the objective of your work at the end of the introduction: ‘Overall, the objective of this study is to …’.
- 82 to 85: I like this idea of explicitly describing very close expressions in the introduction, but I think that it would improve the readability if these expressions were defined in the running text after the first mention.
- Not that I want or ask you to cite my work, but I think that your manuscript is closely related to Bouchard et al. 2024 (https://doi.org/10.5194/tc-18-2783-2024), in particular, in section 4.1.2 where snowpack observations following ROS are thoroughly documented and supported with photos. Based upon the analysis and the hypotheses exposed by Bouchard et al. (2024), how does your study contribute to deepening the understanding of the impact of ROS in boreal forests? I think that this would reinforce the introduction and the first part of the Discussion. Other studies relevant to your work include Teich et al 2019 (https://doi.org/10.1016/j.foreco.2019.01.052) and Bouchard et al 2022 (https://doi.org/10.1002/hyp.14681). Perhaps you can have a look at these.
- 86: I recommend sorting this section (Materials and methods) into two subsections: 1) study site and 2) Experimental protocol. Moreover, this section should include a map of the site. Perhaps, you can move figure S1 to your study site subsection and complement it with an overview of the site location relative to the three weather stations and the town of Khotkovo with a relevant background map.
- 87: Please mention the coordinates of your site (l. 101) here.
- 89 to 91: Please be more specific here. What is the mean winter temperature and precipitation here?
- 91 to 93: How variable are snowpack properties from year to year? Some references are indicated but it would be necessary to mention the information explicitly here.
- 96: at what height are the met data measured?
- 103: Is the site surrounded by topographic features (e.g. hills, mountains) that would alter wind or radiation?
- 105: what height are the surrounding trees?
- 111: Please indicate explicitly the dates of snow pit measurements
- 115. Were snow density and temperatures measured as profiles (at what vertical precision?) or based on the identified snow layers? Please specify.
- 116 to 117. This information should appear later (l. 130) where you describe the related protocol.
- 125 to 128: this should be included in the running text.
- 130 to 135. The description of your protocol here is slightly elusive. I would expect more precision on the morphological characteristics that you aim to quantify. Is it the length of the percolation fingers, their diameter, the tortuosity? Also, as you mention, this approach includes important limitations (l. 134). The uncertainty on the estimation of the density, SWE and contribution to the mass balance should be quantified.
- 136: a reference or a link to the software should be added to your text.
- 136 to 143: Please detail more thoroughly what Scaniverse does and how it is used to identify the percolation funnels. A reference would be helpful here. Does Scaniverse use a objective threshold to determine or is it set by the user? Also, be more specific about the analyses that you perform with the ‘Python libraries’. Maybe the link to Zenodo should be already available.
- 152: As a scalar, wind speed is bounded to 0 m s-1 by definition. Saying that it fluctuated between 0 and 5 m s-1 sounds a bit clumsy. Maybe say that wind speed reached 5 m s-1.
- 155 this sentence is not useful. Please remove.
- 150 (Weather conditions): This whole section includes results that do not belong here as are described the effects of weather on snow stratigraphy (l. 155 to 162; l. 163 to 165; 167- 170). These results should be moved to the following section, and this section 3.1 should be thoroughly reorganized to include only results related to weather conditions: temperature, wind speed, precipitation and snow depth.
- 171 (Figure 1): Where does snow depth measurement shown in Figure 1 come from? Was it measured at one of the weather stations or is the snow depth time series retrieved from your field measurement; I am not sure if this was mentioned. Also, can you explain why the initial snowfall caused the formation of a compact basal layer (l. 157) but Fig 1c shows that snow completely melts out on December 2?
- 175 (Figure 2): I think that this figure should be removed as it repeats what is shown in Figure 1.
- 179 to 180: This was already mentioned in Methods. Please remove.
- 178 (Section 3.2): This section essentially presents how the snow stratigraphy evolves until the ROS of 21 December. It gets a bit hard to follow since all 4 sites are presented at once. Maybe you should specify at which location were measured the lower and upper bounds of snow density ranges that you indicate.
- 183 to 184: Is this what you show on Figure 1. Please specify.
- 198 (Figure 3); Here you only show snow profiles after the ROS despite describing mostly conditions before in the text. It would be insightful to have the profiles (grain type, density, SWE, temp) before and after. Moreover, The density profile should be presented as discretized values as each point is representative of one layer. Then I don’t think it would be necessary to present the profile of SWE, but rather indicate the bulk SWE of the snowpack (before and after ROS).
- 203 to 204: Please review the syntax of this sentence
- 207 to 210: This belongs to methods. Please remove.
- 212 (Figure 4): Were the KDE maps normalized. I find it curious that it reaches 1.
- 226 to227. The standard deviation of the nearest-neighbor distance would be needed here to support this statement.
- 228 to 230: Can you elaborate on the Ripley’s L-Function and the CSR? Maybe it is just me, but this approach does not seem to be meanstream. I think many readers (like me) may get lost here.
- 232 to 234: I feel like this belongs more to the discussion than the results
- 241: Very interesting!
- 255: Please elaborate on that and specify what you mean by ‘directly influenced by tree canopies’
- 272: It would be interesting to show a photo with the analysis from Scaniverse of these elongated flow pathways similar that shown on the left side of Figure 4.
- 275: What is the aspect of the slope?
- 280: Study from Eiriksson et al 2013 (https://doi.org/10.1002/hyp.9666) is of interest here.
- 282 (Sectoin 3.4). In addition to the schematic diagram (Figure 6), It would be relevant to add photographs mentioned earlier (l. 131) in this section.
- 297: Figure 6
- Panel c (forest site) should be removed as this was not supported by field evidence (at least not that is mentioned in the text).
- Indicate clearly on top of each frame which is the meadow and which is the field.
- If you have any photos of the snow pit wall showing the percolation columns (such as Fig. 5 from Bouchard et al. 2024), add it here.
- Can you explain why in Figure 6a, you have dry FC on the same layer as wet FC next to your percolation column? Please also correct FC crystals (the C already stands for crystals). Also specify what do you mean by wet FC, as FC is typically specific to dry snow.
- 309: Can you explain how macropores were formed in the snowpack before it got wet?
- 314: Can you confirm that you have observed more than only one column at each site, and perhaps rephrase? As I read this, it sounds that only one column at the meadow and fields sites was identified and this wouldn’t be enough to draw a generalized conclusion.
- 321: I don’t understand here. Did you assess percolation columns also at the forest site or not? It is confusing.
- 328: have you observed merging of preferential flow pathways at transition layers like Bouchard et al 2024 did?
- 344 (Figure 7): How I interpret this figure is that after some lateral distance water penetrates the snowpack vertically under homogeneous flow rather than preferential flow as is the case for the meadow and field site. Am I correct? If so, this should be somehow highlighted as it would be, in my opinion, an important finding of your work. Also, have you observed from a snow wall dug in the direction of the slope these ladder-like features of percolation or is this what you infer from your three snow pit observations? This should be specified.
- 346 to 362: This should appear before Figure 7 and the corresponding paragraph.
- 366: If liquid water refreezes on contact with snow surface is either because it was freezing rain or because the snow surface was below zero. Looking to Figure 2, We see that the temperature rises above zero before precipitation is recorded within the identified ROS period. Can you elaborate on that and address the role of pre-ROS snowpack conditions on water transport (vertical and lateral). In other word, what one can expect from ROS in the specific locations that you have presented if we were later in the season or in spring, when the snowpack is at 0 C?
- 386: I highly recommend you expand the discussion on limitations as these are numerous in your study. This starts with the fact that your conclusions rely on a single event at only one site. How can this be transposed to other locations under different ROS conditions? Also, your assessment of percolation column characteristics does not rely on any proven protocol and no quantification of uncertainty is provided. This lacks robustness and should be corrected.
- 406 to 410: The low-cost attribute of the photogrammetry approach is emphasized as a strength of this work (also in the abstract) but in fact, the Scaniverse software used in this study is not supported by any reference how can we know that this method was proven and validated? This must be address for an eventual publication.
- 405 (4.3 Broader use and future potential: The only point that is discussed here is the use of low-cost phone-based photogrammetry. However, I don’t think that this is the most important takeaway of your work. In fact, I would expect to discuss how your findings, i.e. the differences observed between locations, can be used in snow, hydrological and land surface models. How can this be integrated into a multi-layer snow models and how to upscale those sub-grid processes at larger modeling scale? I invite the authors to refer to the work of Mazzotti et al (2024) for some ideas (https://doi.org/10.5194/tc-18-4607-2024)
- 451 (Bibliography). Many items in the bibliography are not cited in the text (I have counted 3). Please make sure that the list of references matches the entries the manuscript.
- Except for Fig. S1, none of the figures in the supplement are indicated in the manuscript.
Citation: https://doi.org/10.5194/egusphere-2026-3873-RC2 -
AC2: 'Reply on RC2', Anton Komarov, 25 Sep 2026
We thank the referee for the detailed and constructive review and for recognizing the originality of the observations and the objective of the study. We agree that the manuscript should be framed more clearly as an event-scale case study rather than as a general conceptualization of snowpack response to ROS.
Accordingly, we will revise the Abstract, Introduction, and Discussion to emphasize the limits of generalization from a single natural ROS event. We will also reorganize the Materials and methods section to clarify the study sites, observations, measurements, and the distinction between directly measured characteristics and post-event interpretations.
We will strengthen the treatment of uncertainty by quantifying measurement variability where possible and by discussing the limitations associated with the single-event design, site differences, manual feature delineation, and destructive measurements of individual percolation columns.
The Discussion will be reorganized to compare the observations more explicitly with previous ROS studies, distinguish observations from process interpretations, and discuss implications for multilayer snow, hydrological, and land-surface modelling. Smartphone-based mapping will be retained as a secondary methodological contribution.
We will also address each specific comment individually and revise the corresponding text, figures, and references as appropriate. These changes will improve the methodological transparency, treatment of uncertainty, and positioning of the manuscript as a detailed field case study of spatially heterogeneous snowpack response to a natural ROS event.
Citation: https://doi.org/10.5194/egusphere-2026-3873-AC2 -
AC3: 'Reply on RC2', Anton Komarov, 04 Oct 2026
Dear Reviewer,
Thank you once again for your detailed and specific comments. We are continuing to revise the manuscript. Addressing each individual comment, including revisions to the figures, has taken somewhat longer than expected. However, we plan to finalize our response within the next few weeks.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 165 | 88 | 43 | 296 | 79 | 41 | 35 |
- HTML: 165
- PDF: 88
- XML: 43
- Total: 296
- Supplement: 79
- BibTeX: 41
- EndNote: 35
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Major comments:
This manuscript analyzes the characteristics of snowpack and the distribution of internal liquid water under different surface conditions in the southern taiga based on a single ROS event, which required the authors to invest significant human and material resources. The work deserves recognition. However, judging by the manuscript’s title, its primary aim is to examine the impact of ROS events on snowpack characteristics and the distribution of internal liquid water under different surface conditions. It is undeniable that a single ROS event can provide some insights. However, as shown in Figure 1, this ROS event was a small-scale one, with daily rainfall of approximately 5 mm. If precipitation were to increase to 10, 20, or even 30 mm, and snow depth reached 50 cm or greater, would this phenomenon still occur? Of course, this cannot be studied through natural ROS events—at least not in the short term—because the occurrence of ROS events is random and unpredictable. However, it can be determined through simulated rainfall experiments and model simulations. We recommend that the authors expand upon this manuscript to make its findings more widely applicable. Unfortunately, the current manuscript does not meet the journal’s quality requirements.
Specific comments: