Using a fixed-wing UAS to map snow depth distribution: an evaluation at peak accumulation

De Michele, Carlo; Avanzi, Francesco; Passoni, Daniele; Barzaghi, Riccardo; Pinto, Livio; Dosso, Paolo; Ghezzi, Antonio; Gianatti, Roberto; Della Vedova, Giacomo

doi:https://doi.org/10.5194/tc-10-511-2016

Articles | Volume 10, issue 2

https://doi.org/10.5194/tc-10-511-2016

© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.

https://doi.org/10.5194/tc-10-511-2016

© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 10, issue 2

Research article

|

04 Mar 2016

Research article |

| 04 Mar 2016

Using a fixed-wing UAS to map snow depth distribution: an evaluation at peak accumulation

Carlo De Michele, Francesco Avanzi, Daniele Passoni, Riccardo Barzaghi, Livio Pinto, Paolo Dosso, Antonio Ghezzi, Roberto Gianatti, and Giacomo Della Vedova

Download

Final revised paper (published on 04 Mar 2016)
Preprint (discussion started on 20 Feb 2015)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC C333: 'Comments on the paper by De Michele et al.', Anonymous Referee #1, 07 Apr 2015
- AC C914: 'Answer to reviewer', Francesco Avanzi, 28 May 2015
RC C948: 'Review of “Microscale variability of snow depth using U.A.S. technology” for The Cryosphere', Steven Fassnacht, 03 Jun 2015
- AC C1101: 'Answer to reviewer', Francesco Avanzi, 27 Jun 2015

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Francesco Avanzi on behalf of the Authors (27 Jun 2015) Author's response Manuscript

ED: Referee Nomination & Report Request started (25 Jul 2015) by Philip Marsh

RR by Steven Fassnacht (12 Aug 2015)

Suggestions for revision or reasons for rejection

The authors of the revised version of "tc-2015-16" have done good job addressing both of the reviewers comments. Overall the text reads much better and the Methods are presented much clearer.

The two remaining problems are: 1) the addressing of the reviewer comments and 2) the the use of 12 points to evaluate the dataset. With the first point, the authors make statements such as “We agree with your point of view. We will try to include a wider context about survey uncertainty with respect to the existing literature.” Be specific. What did you do. Tell us explicitly what the changes are. In reading the revised version, I can pick up some of these. Did you just consider what we asked or did you actually address the comment.

For the second point, the authors state in the revised paper: “In the literature, it is known how to determine the number of points that would theoretically be needed to get an estimation of mean snow depth within a certain error range ... On the contrary, no specific rule or common practice exists to determine the minimum number of manual snow depth measurements, over a given area, needed to evaluate whether a given remote sensed technique returns satisfactory performances or not. Although this amount of points allows for a preliminary evaluation of the performances, more data could help to assess the performances of this technique in a more extensive way.” While there is some truth to this statement, they only measured 12 points. This is by no means enough. The paper is interesting enough that it would not be prudent to ask the authors to repeat this effort and “get more points.” So, they should acknowledge that this is too few but that’s what they measured at the time. The should “own it” and move on without vast analysis based on 12 points. Such analysis should be removed. For example, in lines 241-265 they discuss the various interpolation methods for the manual data. This is not necessary. There is enough literature related to interpolation of snow data, so they should state that various methods exist (give a few references), pick one method (kriging) and move forward with that. In section 4.2, the “point evaluation” section could be removed.

In lines 232-240, they degrade the quality of the interpolated UAS-based dataset. There could be references to existing literature (Cline et al., 1998 Hydrol. Proc.; Molotch et al., 2005 Hydrol. Proc.). The component examining the impacts of degradation is reasonable.

The authors have taken the review comments to heart, especially for improving the figures. Figure 6 is an interesting and relevant addition to the paper. I don’t know what Figure 9 adds, and Figures 10 and 11 should be stacked and clarified. The y-axes are not clear and when these go to print the lines will be hard to see.

The “Outlook” section at the end is a good touch.

Hide

RR by Anonymous Referee #3 (13 Aug 2015)

RR by Davide Bavera (20 Aug 2015)

Suggestions for revision or reasons for rejection

General Comments:

The paper presents a very promising application for snow measurement and SWE assessment in the near future that could be quite cheap and easy to retrieve. The topic is relevant for the journal and interesting for scientific and technological development, for hydrological aims, and for other applications (e.g. avalanche monitoring and safety).
At the same time, anyway this contribution is week in terms of comparative analysis and validation of its results.

Some comparisons should have been made against one of the already tested distributed techniques (e.g. laser scanner) to be more effective and relevant. Or, at least, against a validated physical based model or even using a much more dense set of manual measurements. This at least to have evidence of the expected effect of local topography and provide a validation of the method a little bit more reliable and significant.

To improve reliability and relevance of the measures and methods it is advisable to define and put in place some snow poles, properly and precisely located with GPS. These can be also used referencing images afterwards providing common points, and could give more information than probes, for time depletion curves, accumulation and melting processes. Finally, in any case I would add a great number of measure with snow probes in the peak accumulation survey as already suggested by other reviewers.

As the authors suggest, a more detailed and exhaustive comparison with existing techniques is advisable, and a greater number of points for snow probe measurement is required.

Deductions described in paragraph 4.3.2 appear not to be significantly innovative and quite week and not really straightforward. They sound to be not really convincing. Dataset is and remain quite poor and this statistics analysis is not giving a real plus, in my opinion.
The most informative analysis in this paragraph, in my opinion, is the part regarding standard deviation and CV (fig. 11). I would summarize or remove the other parts.

Paragraph 4.3.3 assumes without any validation that U.A.S. output for snow volume is the correct one. Even if the analysis then is correct and appropriate, a rigorous validation against a spatial already tested technique is missing and should be included to provide strengthens to the assertions.

Outlook section is essential to define the awareness for limitation of this research. In any case I would recommend to strengthen the sample data and/or the comparison with other distributed (remote sensing data or physical model) dataset.

Specific Comments:

Title: I would remove resolution info in title

l. 10: manual measurements: how? Where? Provide more details. As already said you should have collected more than 12 points

l. 17: add results of point values also in percentage

l. 21: include a short motivation about snow volume difference

l. 104: in my opinion the survey should has been made for two or three years to have more information

l. 109: include more details about 12 points measurements

l. 121: change "since it is an elevation that guarantees" with "since the elevation guarantees". I would stress more clearly that it is important that the study area has NO permanent snow.

l. 275: in order to better verify the quality of the surveyed data it would be better to compare the U.A.S. DSM with an already validated one, computing some statistics of the differences between the two DSM: please show results of the comparison you made.

l. 315: snow probe measurements should be more numerous and possibly located in flat zone, local maxima and local minima of the topography, in order to get the greatest variability

l. 340: "Basing on these results, a spatial resolution of 20 cm seems to be the trade-off between the number of pixels considered (i.e., computational time) and the description of the snow micro-topography for the considered area": this deduction is not straightforward. How do you evaluated 20 cm as the bets compromise? Rougher resolutions have not been investigated and an evaluation index is missing.

l. 345: "While the three maps considered in the previous section...": if you talk about these maps you should show them in figures

l. 349: "...the larger cell size, the lower degree of detail..." this assertion is quite obvious, please rephrase the sentence

l. 445: I would use "vehicle" or something similar instead of "support"

tab. 1: it seems that absolute and percentage (please add this information to the table) difference are greater for lower snow depth. This is reasonable but should be more in detail investigated and commented, and, also for this aim, much more measurements would be required.

tab. 2: negative values in Hmin? Explain the meaning or remove.

tab. 3: result clearly show that the snow probe sample size and location is not suitable to assess snow volume, assuming that U.A.S. value is reliable. Please consider to significantly improve the snow probe sample, and detail comment section about this.

fig. 3, 4, 5 and 7: make the upper layers a bit transparent in order to make visible the underlying map.

fig. 9: please consider to change the color map in order to better highlight where the difference is close to zero, where you have overestimation and where you have underestimation.

fig. 10 and 11: please consider to use different symbols instead of colors to allow black and white print and add the legend in the plots.

Grammar, typing, etc.: please check time of the verbs (past and present are used not in coherence) and improve readability of numbers (thousands, etc.)

Hide

ED: Reject (19 Oct 2015) by Philip Marsh

ED: Publish subject to minor revisions (Editor review) (20 Jan 2016) by Philip Marsh

AR by Francesco Avanzi on behalf of the Authors (02 Feb 2016) Author's response Manuscript

ED: Publish as is (12 Feb 2016) by Philip Marsh

AR by Francesco Avanzi on behalf of the Authors (12 Feb 2016)

Short summary

We investigate snow depth distribution at peak accumulation over a small Alpine area using photogrammetry-based surveys with a fixed wing unmanned aerial system. Results reveal that UAS estimations of point snow depth present an average difference with reference to manual measurements equal to -0.073 m. Moreover, in this case study snow depth standard deviation (hence coefficient of variation) increases with decreasing cell size, but it stabilizes for resolutions smaller than 1 m.