the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 23 Aug 2024
Aspect Controls on the Spatial Re-Distribution of Snow Water Equivalence in a Subalpine Catchment
Abstract. Quantifying subalpine snowpack parameters as they vary through time with respect to aspect and position on slope are important for estimating the seasonal storage of snow water resources. Snow depth and density are dynamic parameters that change throughout the progression of the accumulation and melt periods, with direct implications on the distribution of Snow Water Equivalence (SWE) across a landscape. Additionally, changes in density can infer physical processes occurring within the snowpack such as compaction, liquid water pooling, and lateral flow. This study measures snow depth and density throughout a 0.25 km2 watershed in northern Colorado USA using L-Band (1.0 GHz) Ground Penetrating Radar (GPR) and coincident depth probing. GPR snow densities were calibrated using bulk densities from snow pits and a SNOTEL station. A physical snowpack model, SNOWPACK, with input from local Remote Automated Weather Station and SNOTEL station produced models of snow depth, snow density, and liquid water content (LWC). The model simulations indicate mid-winter melt events produced LWC on the south aspect that are less present in the north aspect and flat areas. These midwinter melt events resulted in the lateral flow of LWC downslope, and the redistribution of SWE as observed in GPR surveys. Further observations show a steady increase of soil moisture throughout the winter in the flat terrain and ice layer formation on the south aspect snow pits during mid-winter surveys. Other key observations include pooling of liquid water at the base of the north aspect during the later spring season melt phase evidenced by pit observations and GPR transects. We further develop a conceptual model for the aspect controls on the distribution and movement of SWE during the winter and spring seasons. In summary, mid-winter melt events are observed on south aspects, causing a redistribution of SWE downslope while spring melt brings liquid water pooling at the base of north aspects. These differences in snowmelt dynamics based primarily on aspect, providing important processes to consider for spatially and temporally extensive SWE measurements moving forward.
- Preprint
(1958 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-2364', Anonymous Referee #1, 09 Oct 2024
In the study 'Apsect Controls on the Spatial Re-Distribution of Snow Water Equivalence in a Subalpine Catchment', the authors investigated how slope aspect influences snow accumulation and melt dynamics using ground penetrating radar (GPR) transects and snow pit and snow core density measurements during the water year 2022-2023 in a small catchment in Colorado, USA. The authors also used the SNOWPACK snow model to support their observations and further develop a conceptual model of water redistribution in the snowpack based on aspect. Overall, this study provides interesting results using different instruments and tools (GPR surveys, snow pits, weather stations and a multi-layer snow model). I acknowledge the authors' efforts in combining these elements to propose a conceptual framework to better understand how aspect may exert a control on the evolution and (re)distribution of snow water equivalent (SWE). However, the proposed conceptual model is based on indirect observations and modelling results rather than on field evidence, which should be clearly stated. I also believe that the manuscript would benefit from a more straightforward storyline and some elements such as the influence of forest canopy on SWE should be detailed. After addressing these issues and some additional minor and technical comments, I am confident that this paper will be a good and relevant scientific contribution to snow research and will provide insights for future studies.
Major comments.
Structure of the introduction
The introduction needs to be clarified so that the reader understands the relevance of the study. Here are some suggestions:- The first two paragraphs would benefit from being restructured into one and generally rephrased and shortened.
- The third paragraph (l.47 to l.73) is too long and goes in all directions. I would recommend the author to split this paragraph into two, based on the description of methods (l. 47 to 60) and landscape control of snowpack properties (l. 61 to 73).
- The last paragraph should be completely revised. Lines 85 to 91 should be moved earlier in the Introduction. I would recommend that authors make links to earlier parts of the introduction to emphasise the relevance of their work. I also strongly suggest that a research aim for this work be clearly defined (which is not the case at present).
Modeling setup
Several questions remain about the modelling setup.- What parameters did the authors use for each simulation (north, south flat)?
- Were soil layers defined in these simulations? If not, is there a reason for this, since the authors have a description of the soil (see lines 101 to 104)?
- There are several parameters in SNOWPACK that are site-dependent or have to be chosen arbitrarily by the modellers. What values did the authors use in their simulations for these parameters?
- I understand that the canopy was taken into account for the flatland simulations. This implies that several other parameters have to be specified. What values did the authors choose?
I think adding an appendix with the different keys enabled in the model and the parameters used for each simulation would be a clever way to answer these questions. Consider also explaining any arbitrary choices and how site-specific values were obtained.
Influence of the canopy
One of my main concerns with this article is that the influence of canopy cover on the spatial distribution of SWE is not adequately addressed. While canopy control is presented in the Introduction (l. 66 to 71) and some results are interpreted based on the canopy in the Discussion (l. 294-298; l. 325-334), the role of vegetation is not presented in the Results section. While I respect the authors' decision not to make this the main focus of their paper, I think the article would benefit from consistent treatment of the influence of canopy on snow redistribution alongside aspect control. Please consider including this in a revised version of your manuscript.Limitations
The fact that not every interpretation is based on field evidence is not critical. However, I would suggest that the authors include a section on the limitations of their study. This would help the reader to better contextualise some of the analyses, especially with regard to SWE redistribution processes through the snowpack.Minor comments.
- l. 16. Explicitly mention the use of snow pit and soil moisture monitoring measurements in the abstract.
- l. 32-33. I do not think this sentence is necessary. Please remove.
- l. 33-36. This sentence is difficult to understand. Please rephrase and break it down into two sentences.
- l. 42-46. I am not sure if I understand this sentence correctly or if it is necessary for the general understanding of your study. Please rephrase or clarify this idea.
- l. 57-60. Please consider breaking it down into two sentences.
- l. 61. I think starting a new paragraph here would improve the readability of the introduction.
- l. 66-68. You mention the energy balance, but then refer more to the mass balance of the canopy (e.g. accumulation by canopy, interception). Perhaps you should just mention that the canopy changes the energy and mass balance of the snowpack.
- l. 74. Consider specifying the 'bulk' snow density here. The distinction is particularly important as you go on to present detailed density profile measurements (Fig. 6). I would also consider adding a few words on how snow density at the layer scale varies with landscape characteristics.
- l. 77-80. Why is the derivation of snow density from permittivity given for dry snow only? A few words about this method applied to wet snow would be relevant.
- l. 82. Please clarify the meaning of ‘spatial relationships’.
- l. 87. I am a bit uncomfortable with ks being the symbol for the dielectric permittivity of snow. ks often refers to the thermal conductivity of snow. Please consider using the symbol 'ε' for permittivity.
- l. 90. Please delete the following: ‘being dragged as fast as a surveyor can traverse the snow’.
- l. 90. What is ds? This variable has not yet been defined.
- l. 100. Please specify the historical period of the measurements.
- l. 103. Do you have the average thickness of the litter? If so, please specify.
- l. 110-119: Please consider shortening the details of how the DEM and canopy height models were developed.
- l. 114. I understand the meaning of the word ‘canopied’, but as this term is quite uncommon, it distracts the reader from the text. Consider using another term.
- l. 117. Please check and correct the end of this sentence.
- l. 118-119. I do not think this sentence is necessary.
- Figure 1.
- Why is north pointing to the left? I think it would be better to rotate your map 90 degrees and make it poiting upward. I am not sure that the orientation of 1a is the same as 1(b to e), please check. Consider adding the river to figure 1a.
- 1 b and Fig. 1d could be combined into one figure using elevation lines.
- 1c could be removed. If the authors decide to keep it, please indicate how shortwave radiation was calculated.
- l. 121-123. Please rephrase. It took me a few reads to understand the sentence.
- l. 126. Please indicate the exact start and end dates of the data collection.
- l. 131-133: Please revise these two sentences. It seems that some words are missing...
- l. 135. Why did you use two different systems? And how might this affect your results?
- l. 137-142: Please consider adding a table of snow pit measurement dates, indicating which density measurement method (wedge cutter or tube) was used on which date.
- l. 142. Please indicate how water ponding and ice lenses were identified. Perhaps a photo of a snow pit experiment (if you have one) would be relevant here.
- l. 145. I would remove Figure 2 from the manuscript.
- l. 147. Please add a few words about ReflexW.
- l. 147-163. I really appreciate this paragraph, which is fluent and easy to read. I think a conceptual figure of the multi-step data processing method would be nice. Please consider replacing Figure 3 with this conceptual figure.
- l. 167-175. I get quite confused with ds and ks. Defining ds first would definitely help, but still. This part with the equations is a bit messy. Please check that the correct variables are used and described. Please also include the number of each equation.
- l. 180-181. These two sentences should be merged into one.
- l. 182-183. This sentence should follow the description of the data provided by the SNOTEL and RAWS stations.
- l. 191. Please mention that redistribution (e.g. by wind or canopy unloading) is neglected.
- l. 198. 2023 water year? Consider adding a label on the x-axis of the plot instead.
- l. 200. Could you add a sentence explaining why SNOWPACK was used instead of another model?
- l. 204-205. This is not exact. Please be more specific about how SNOWPACK creates, removes or merges snow layers.
- l. 206. A clearer explanation of the liquid transport processes could be given here. See Wever et al. (2014 - https://doi.org/10.5194/tc-8-257-2014).
- l. 207. In fact, SNOWPACK relies on fundamental physical principles to simulate snow metamorphism. Please remove the statement that it has ‘a unique empirical scheme’.
- l. 235-237. Can this be confirmed by any snow pit observations?
- Figure 5. While I appreciate the effort put into this figure, I think it could be simplified. The way the figure is presented makes it difficult to compare results from different sites. Also, in section 3.1 of the the text, the frames (a, b, c ...) are not presented in any order, which makes it confusing. I would suggest a typical side-by-side plot where we can more easily compare the north-facing slope, the south-facing slope and the flat terrain.
- l. 251. As snow pit observations were not systematically performed during your field surveys, I would recommend listing each snow pit date in a table (perhaps in the method section).
- l. 257-259. That is an interesting observation. Could you elaborate?
- Figure 6: Please increase the size of the axis labels. Consider also using a colour gradient to display density profiles (see Fig. 3c-d from Bouchard et al. (2022 - https://doi.org/10.1002/hyp.14681) as an example). This would allow each profile to be shown on the same frame and would make them easier to compare.
- As a general comment, be sure to follow a same order of presentation of the results (e.g. 1. flat, 2. south, 3. north) in the different sections where you refer to them.
- l. 263. Although this is not the objective of the study, I think it would be interesting to compare the simulation results for snow density with your snow pit observations. This would give a better idea of how the model performs at your site. Consider adding this analysis.
- l. 272. The difference in peak SWE is huge! I think this needs to be highlighted and explained.
- l. 274-275. Is this based on volumetric water content (Figs. 6c-d-e)? I think the surface runoff simulation would be interesting here. Consider adding them to Figure 7.
- Figure 7. Units and date formats should be consistant with other figures (Figs. 4 to 6).
- l. 283-284. In fact, ponding of liquid water at the base of the snowpack was not demonstrated by your results, but rather suggested by simulations and SWE observations. However, evidence of ponding could be provided by snow pit observations. If you have such observations of ponding at the base of the snowpack, consider adding them. Otherwise, please revise the wording of this sentence.
- l. 288. Just to be sure, by observational data, do you mean the SNOTEL station measurements?
- l. 298-300. The comparison with the northern aspect remains speculative as there were no wind speed measurements taken there.
- l. 302-304: Have you applied any wind undercatch corrections to the forcing precipitation?
- l. 309-310. This response may be enhanced by lateral flow over ice layers in the snowpack. See Eiriksson et al. (2013 - https://doi.org/10.1002/hyp.9666).
- l. 313-314 and Figure 8. This should be moved to the Results section.
- l. 316. Do you have any temperature observations from your snow pit observations (even once) to support this?
- l. 342. Can you elaborate on the prevalence of hydraulic barriers in the northern aspect snowpack rather than in the southern aspect snowpack?
- Figure 9. This conceptual figure is interesting, but it is not based on field evidence. This should be clearly stated in the text.
- I recommend that the authors compare their results with those of Mazzotti et al. (2023 - https://doi.org/10.5194/hess-27-2099-2023)
- l. 380-381. This has not been directly observed and remains a hypothesis. I would refrain from drawing conclusions from this.
- l. 384. Please add a few words on how these results would differ in different locations/climates. Please also add some concluding remarks on how the results of this work can improve our global understanding of snow in complex terrain and provide guidance for future research.
Technical comments.
- l. 12, 20, 24 and so on… Please consider using the term “ponding” instead of “pooling” throughout your manuscript.
- l. 12. This study measures --> In this study, we measured.
- l. 15. input --> inputs
- l. 16. models --> simulations
- l. 16. missing word (that?)
- l. 21. (snow) pit.
- l. 31. ‘Regional distributions in SWE also impact ecosystem services through surface albedo, effectively cooling earth surfaces and regulating climate’. It took me a few reads to understand this sentence. I recommend the following change: ‘Regional distributions in SWE also impact ecosystem services through surface albedo, which effectively cools Earth’s surfaces and regulates climate.’
- l. 41. measure --> estimate
- l. 47. I do not get what you mean by “snow cover” being a snowpack properties.
- l. 87. Please include the year of that reference
- l. 97. Please include the year of that reference
- l. 98. Please indicate that masl means meters above sea level
- l. 100. Please verify the format of the date.
- l. 127. were --> was
- l. 153, un-necessary --> unnecessary (?)
- l. 225-226. Please revise the syntax of this sentence
- l. 265-266. Please, revise this sentence.
- l. 266-267. Please indicating Fig. 7a-b only once.
- l. 294. doesn’t --> does not
- l. 318. Please remove “and requires further research in the future”.
Citation: https://doi.org/10.5194/egusphere-2024-2364-RC1 -
AC1: 'Reply on RC1', Ryan Webb, 21 Oct 2024
Thank you for the thorough review of our manuscript. Attached is a response document with responses to each of your comments and proposed revisions to address each. Please let us know if you have further questions/comments for discussion prior to the closing of the open discussion period.
-
RC2: 'Comment on egusphere-2024-2364', Anonymous Referee #2, 27 Oct 2024
This study examines the influence of aspect and slope position on snowpack parameters i.e., depth, density, and liquid water content (LWC), within a subalpine watershed in Colorado, USA. The variations of these parameters are evaluated using GPR, in situ stations, snow pits and SNOWPACK modeling. The study found that mid-winter melt events predominantly affect south-facing slopes, triggering later flow of LWC downslope and the redistribution of SWE. Additionally, ice layers develop on south-facing slopes during mid-winter periods. Flat terrain exhibits a steady increase in soil moisture throughout the winter. In contrast, as spring progresses, north-facing slopes witness the pooling of liquid water at their base.
The findings underscore the importance of considering aspect and slope position when estimating snow water resources. However, many conclusions are based on qualitative reasoning and are not always support by the collected field evidence. While the snow modeling community is undoubtedly moving towards better representation of complex snow redistribution and melting processes, this paper does not provide sufficient quantitative evidence to significantly advance our current understanding of snow dynamics. If the authors intend to maintain a qualitative and conceptual approach, the manuscript should be retitled to reflect this focus. Additionally, a dedicated section should be included to address the study limitations. For instance, the paper could discuss why factors such as wind, canopy, terrain roughness, and eventually gravitational transport were not explicitly considered in this analysis.
Major comments.
- While I appreciate the complexity of organizing extensive snow campaigns and the integration of various tools like GPR, snow pits, and SNOWPACK, I'm uncertain about the optimal utilization of GPR in this study. While GPR can efficiently survey transects, its application here seems to be limited to average this information to a single-point observations (derived from averaged TWT and snow depth along the transect). The potential uncertainty associated with this approach is not explicitly addressed, and it appears to be significant. Additionally, GPR limitations in wet snow conditions and its inability to provide detailed snow layering information, particularly regarding ice lens formation or wind redistribution, makes the use of GPR difficult to justify in this work. Furthermore, the absence of radargrams as supplementary materials, which is an interesting data per se, hinders reproducibility and future works.
- The paper introduces the canopy influence as a key factor affecting the energy balance (L66 on), yet the specific role of canopy within the study domain remains unclear. While LiDAR data is mentioned and depicted in Figure 1e, its utilization in the analysis is not explicitly detailed. The discussion on canopy effects often lacks specificity, relying on generic considerations rather than relate to the specific test site. Similarly, the approach to estimating snow density from GPR data is confusing. The introduction suggests that density is generally considered uniform and that GPR can provide spatialized accurate measurements (L74 on). However, the subsequent averaging of density along transects contradicts this assumption. It would be beneficial to see a comparison of the radargrams, also at a qualitative level, before averaging them (this may further support the conceptual model of Fig 9). Additionally, the absence of uncertainty quantification in the results section hinders the interpretation of comparisons and the reliability of conclusions. I suggest addressing these points, such that the paper can strengthen its scientific rigor and provide a more comprehensive understanding of the complex interactions between canopy, topography, and snow processes.
Detail comments
L14 From Sec 2.3. it is not clear how the calibration of GPR snow density is done using snowpits and SNOTEL stations.
L23 This assertion seems to be limited to the particular characteristics of the study area and may not generalize to other conditions.
L75 Typically, bulk snow density is measured using a federal tube or within snow pits by summing the density derived by smaller volume tubes (or triangular prisms), as described by Kinar and Pomeroy, 2015.
L91 Snow depth can vary significantly, even over short distances, due to the rugged and heterogeneous nature of alpine terrain. This variability, combined with the small area sampled by a probe, highlights the importance of quantifying uncertainties in snow density estimates. Generally an average of N measurements should be done.
L92 If the primary focus of the research is to investigate the impact of aspect and slope position on snowpack dynamics, a thorough justification is required to explain why factors such as wind, canopy, terrain roughness, and gravitational transport were not explicitly considered in the study, especially given their potential influence on snow distribution and melt.
Fig 1a please rotate it consistently with the other figure (i.e., North up)
L162 Please explicitly state that, as reported in Webb & Mooney 2024c, TWT is calculated as an average value.
L170 the equations must be numbered.
L175 Please provide a method for calculating the uncertainty associated with the TWT measurements. Given the potential for significant error propagation due to small denominator values, a rigorous uncertainty analysis is essential.
Section 2.4 how the SNOWPACK free parameter has been calibrated?
Figure 5 is difficult to interpret. A simpler, more traditional visualization would improve the comparison of differences between the data.
Figure 6 please report the uncertainty for all the measurements.
L287 “unusual results” respect what?
L305 “model weakness”? Can you better elaborate the sentence?
L308 Can you better justify this sentence showing the evidence of this mechanism?
L 333 Why “unrealistic”? Can you better elaborate it?
L 354 The answer to the main research question of the paper is answer considering only the melting. So, the melting was the focus of the research?
Figure 9. This conceptual figure is interesting, but it is not based on field evidence. This should be clearly stated in the text.
L 367 I suspect that Dingman simplified his modeling to a homogeneous snowpack. While the four-phase model remains valid for individual homogeneous layers, additional complexity is necessary to accurately represent real-world snowpacks (which however is made up of different homogeneous layer, possibly at different phase).
L371 Given the significant spatial variability in snow depth, particularly in complex terrain, it is challenging to believe that traditional probing methods can accurately capture these variations without averaging N measurements and without a rigorous uncertainty analysis.
L374 and conclusion: So this is only a study on the energy balance and not on snow redistribution processes?
As a final note, while there are no explicit publisher guidelines against self-citation, it is generally advisable to minimize excessive self-referencing. For instance, the accurate prediction of LWC by SNOWPACK could be supported by citing previous studies (as done in the current self-cited works) that provide also detailed information about the model details, which is not developed by the authors.
The References section is difficult to read due to the lack of spacing between entries. Additionally, some references appear to be formatted incorrectly e.g., L87 Clark et al. should be Clark et al., 2015.
Kinar, N. J. and Pomeroy, J. W.: Measurement of the physical properties of the snowpack, Rev. Geophys., 53, 481–544, https://doi.org/10.1002/2015RG000481, 2015.
Citation: https://doi.org/10.5194/egusphere-2024-2364-RC2 -
AC2: 'Reply on RC2', Ryan Webb, 08 Nov 2024
Thank you for the comments and thorough review of the manuscript. I appreciate the time you spent reading and providing constructive feedback on our work. Please find attached a document with initial responses to all of your comments. Please let me know if you would like to further discuss any of the responses.
Data sets
Dry Lake Watershed, CO Ryan Webb http://www.hydroshare.org/resource/4aff38a0cbb24456be4e99987e808abb
Dry Lake Transect Snow Depth 2023 Ryan Webb and Kori Mooney https://doi.org/10.4211/hs.1347210139a945048a1a3ecd93f81dd2
Dry Lake Observed Density 2023 Ryan Webb and Kori Mooney https://doi.org/10.4211/hs.8e87b814ce7541e68f9c2a12a9882c09
Dry Lake Ground Penetrating Radar TWT 2023 Ryan Webb and Kori Mooney https://doi.org/10.4211/hs.b84c4fe4a4d04e77ade1bbae4a0c74f3
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 224 | 78 | 40 | 342 | 5 | 3 |
- HTML: 224
- PDF: 78
- XML: 40
- Total: 342
- BibTeX: 5
- EndNote: 3
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1