the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 05 Dec 2024
Leveraging Citizen Science, LiDAR, and Machine Learning for Snow Depth Estimation in Complex Terrain Environments
Abstract. The majority of the water supply for many Western US states is derived from seasonal snowmelt in mountainous regions. This study aims to address gaps in basin-scale snowpack modeling by using a multi-step, Gaussian-based machine learning model to generate rapid, high-resolution, snow depth estimates at minimal cost by combining citizen-science snowdepth observations with static LiDAR terrain features collected at a single snow-free date. We focus on reducing personnel danger by modifying the algorithm to minimize the exposure of field sample collectors to avalanche-prone terrain. Using snow observations taken solely within a subbasin (∼9-km2) of a larger basin (∼70-km2), a basin-scale snow depth estimate is modeled for a given date throughout the snow season. Results show that a small number of observations (i.e., 10) within a subbasin can realize snow depth across the greater basin with high accuracy, with a root mean squared error (RMSE) of 0.37 m, and Kling-Gupta efficiency (KGE) of 0.59 when compared to the true snow depth distribution. We test the universality of the algorithm by modeling multiple subbasins of differing spatial characteristics and find similar results. The algorithm shows consistent performance across subbasins with varying spatial characteristics and maintains accuracy even when highrisk avalanche areas are excluded. This method exhibits a potential for citizen-scientist data to safely provide seamless modeled snow depth across different spatial ranges in snow-covered basins.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(2801 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2801 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-3545', Anonymous Referee #1, 04 Jan 2025
Review on the manuscript titled, “Leveraging Citizen Science, LiDAR, and Machine Learning for Snow Depth Estimation in Complex Terrain Environments” by Liljestrand et al.
The study tried to characterize and model the snow distribution using the Gaussian Process Regression (GPR) method and the Gaussian-based machine learning model (GMM). GMM and GPR method application for the Lidar observed snow distribution seems novel although they were tested using only one snapshot snow distribution. The land surface characterization in Figure 2 is nice, and the finding of elevation variability requirement for this method is interesting. However, since the transferability of this method may still be arguable, I recommend “major revision” for this review cycle for clarifications and further possible improvements. I have a few major points listed below:
- The presentation of the data used in this study should be improved. The observed snow distribution by the airborne Lidar may be visualized and presented somewhere in the manuscript, perhaps instead of Figure 3. It will be informative for readers to see the variability and the extent of the dataset.
- Also, it is unclear when the LiDAR data collected. The snow distributions are highly dependent on season and year. From the snow distributions (Figure 9), I speculate that it must be late spring. Moreover, observation dates of the in-situ snow depth survey must be presented as well. Were they exactly same day? How good were they? Were they (field data vs. Lidar) consistent each other? It is unclear how the authors use actual field measured data. I would suggest adding a data list table.
- The assumption for the methodology must be further clarified. Based on my understanding, Gaussianity in local snow distribution is required while it may not be true. I recall a recent publication in the same journal (TC) discussing non-Gaussianity of snow distribution (https://tc.copernicus.org/articles/18/5139/2024/). Assumption of local Gaussianity may be addressed in the limitation statement in the discussion as a reminder.
- I understand that there was no improvement by increasing sample number from 10 to100. It would be more useful if the author could quantify the ideal snow data point density (for instance, # of data point per unit area, perhaps). I understand it may be beyond scope of this study while lacking statement on potential transferability made this work just a case study based on single instantaneous snow distribution, which is rather weak.
Specific/minor point:
It is good to define the variables in the equations (2 through 4) as physical quantities (e.g. x = snow depth). Capital sigma (=covariance?) may be avoided because you use “summation” as same symbol.
Citation: https://doi.org/10.5194/egusphere-2024-3545-RC1 -
AC1: 'Reply on RC1', Dane Liljestrand, 20 Feb 2025
Thank for your thoughtful and constructive feedback on our manuscript. We appreciate the time and effort invested in reviewing our work and have implemented several revisions to enhance the clarity, transparency, and accuracy. We hope the revisions adequately address the comments and remain open to further suggestions or questions. Please see the attached pdf containing a detailed response to each comment.
-
RC2: 'Comment on egusphere-2024-3545', Anonymous Referee #2, 22 Jan 2025
-
AC2: 'Reply on RC2', Dane Liljestrand, 20 Feb 2025
Thank you for your thoughtful and constructive feedback on our manuscript. We appreciate the time and effort invested in reviewing our work and have implemented several revisions to enhance the clarity, transparency, and accuracy. We hope the revisions adequately address the comments and remain open to further suggestions or questions. Please see the attached PDF of a detailed response to each comment.
-
AC2: 'Reply on RC2', Dane Liljestrand, 20 Feb 2025
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-3545', Anonymous Referee #1, 04 Jan 2025
Review on the manuscript titled, “Leveraging Citizen Science, LiDAR, and Machine Learning for Snow Depth Estimation in Complex Terrain Environments” by Liljestrand et al.
The study tried to characterize and model the snow distribution using the Gaussian Process Regression (GPR) method and the Gaussian-based machine learning model (GMM). GMM and GPR method application for the Lidar observed snow distribution seems novel although they were tested using only one snapshot snow distribution. The land surface characterization in Figure 2 is nice, and the finding of elevation variability requirement for this method is interesting. However, since the transferability of this method may still be arguable, I recommend “major revision” for this review cycle for clarifications and further possible improvements. I have a few major points listed below:
- The presentation of the data used in this study should be improved. The observed snow distribution by the airborne Lidar may be visualized and presented somewhere in the manuscript, perhaps instead of Figure 3. It will be informative for readers to see the variability and the extent of the dataset.
- Also, it is unclear when the LiDAR data collected. The snow distributions are highly dependent on season and year. From the snow distributions (Figure 9), I speculate that it must be late spring. Moreover, observation dates of the in-situ snow depth survey must be presented as well. Were they exactly same day? How good were they? Were they (field data vs. Lidar) consistent each other? It is unclear how the authors use actual field measured data. I would suggest adding a data list table.
- The assumption for the methodology must be further clarified. Based on my understanding, Gaussianity in local snow distribution is required while it may not be true. I recall a recent publication in the same journal (TC) discussing non-Gaussianity of snow distribution (https://tc.copernicus.org/articles/18/5139/2024/). Assumption of local Gaussianity may be addressed in the limitation statement in the discussion as a reminder.
- I understand that there was no improvement by increasing sample number from 10 to100. It would be more useful if the author could quantify the ideal snow data point density (for instance, # of data point per unit area, perhaps). I understand it may be beyond scope of this study while lacking statement on potential transferability made this work just a case study based on single instantaneous snow distribution, which is rather weak.
Specific/minor point:
It is good to define the variables in the equations (2 through 4) as physical quantities (e.g. x = snow depth). Capital sigma (=covariance?) may be avoided because you use “summation” as same symbol.
Citation: https://doi.org/10.5194/egusphere-2024-3545-RC1 -
AC1: 'Reply on RC1', Dane Liljestrand, 20 Feb 2025
Thank for your thoughtful and constructive feedback on our manuscript. We appreciate the time and effort invested in reviewing our work and have implemented several revisions to enhance the clarity, transparency, and accuracy. We hope the revisions adequately address the comments and remain open to further suggestions or questions. Please see the attached pdf containing a detailed response to each comment.
-
RC2: 'Comment on egusphere-2024-3545', Anonymous Referee #2, 22 Jan 2025
-
AC2: 'Reply on RC2', Dane Liljestrand, 20 Feb 2025
Thank you for your thoughtful and constructive feedback on our manuscript. We appreciate the time and effort invested in reviewing our work and have implemented several revisions to enhance the clarity, transparency, and accuracy. We hope the revisions adequately address the comments and remain open to further suggestions or questions. Please see the attached PDF of a detailed response to each comment.
-
AC2: 'Reply on RC2', Dane Liljestrand, 20 Feb 2025
Peer review completion
Journal article(s) based on this preprint
Data sets
Franklin Basin, UT, USA Snow-On LiDAR Dane Liljestrand, Bethany Neilson, and Carlos Oroza https://doi.org/10.4211/hs.34ce22e4df10463cb053bb63d19d6672
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 313 | 97 | 22 | 432 | 15 | 37 |
- HTML: 313
- PDF: 97
- XML: 22
- Total: 432
- BibTeX: 15
- EndNote: 37
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Dane Liljestrand
Ryan Johnson
Bethany Neilson
Patrick Strong
Elizabeth Cotter
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2801 KB) - Metadata XML