UAV Video-Based Approach to Identify Damaged Trees in Windthrow Areas

Round 1

Reviewer 1 Report

 

Interesting, concise and clear article regarding the possible use of full-motion video with UAS during identifying damaged trees in a windthrow area.

Detailed comments:

Line 31: use different words than in the title of the MS to get more visibility of your research for example: downed trees, salvage felling area etc.

Lines 78-82: please include information on how much volume was damaged on that occasion? Also, how mush volume on your study area.

Line 89: please add A.sachalinensis (F. Schmidt) Mast.

Line 112: is the RTK accuracy of 1 cm provided by your national station service?

Lines 114-120: please include the volume of the damaged trees in your study area.

Line 191: please give information on the diameter at breast height of the downed trees (snapped trees if possible). Why did you choose a buffer width if 0.25m?

Line 160/ figure 4.: you have stated that deadwood was classified as a non vegetation (line 160), but how can you be sure (since you have fallen/snapped trees from a 17 year old storm) what is non vegetation and what are parts of two other classes. Please explain this in more detail.

Line 289: your research was in December, so maybe this first sentence is not the best way to start the Discussion even though is it true.

Line 349-351: rephrase to”…deadwood, which remain in the stand longer and play an important role in the forest ecosystem.”

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The Authors have explored the potential application of UAV Video-Based Approach to identify damaged trees in a forested area affected by natural disturbance.

The topic is interesting and this approaches may be, in the upcoming future, the new forestry way to deal with problems related to the inaccessibility of those areas. In order to be published in the Journal, this paper needs to be improved in some parts.

1 – The limitations of SfM techniques are well known in the literature and the difficulties related to its employment are the ones expected, I would like to see more references about it in similar case studies.

2 - The choice to exploit FMV techniques is a good way but this methods needs to be properly described in the introduction section, being it too superficial (rows 61-68). The novelty of the work must be highlighted in comparison with other papers with the same topic, even if the applications are different.

3 – raw 112, add RMSE or, at least, an accuracy value. This comment must be repeated for each statement regarding a “centimeter level” of accuracy when it's reported in the paper.

4 – Workflow figure is clear enought  but it needs to be improved as style, trying at least to enhance the main steps of your work. In general, the figure quality must be improved. Just to make two examples, the screen capture of figure 3 (a and b) has low quality and the figure 1c should be substituted with the output obtained from the SfM techniques so that the quality of figure will certainly increase.

5 – For what concerns rows 131-137 the statement “some adjustments had to be done according to the parameters supplied by ESRI’s support team” must be better explained, adding a reference or trying to explain which and how the parameters where modified. I would like to see some references about the “ESRI’s recommendations”.

6 – The results section should be rewritten in a more readable way. Results are consistent, but they are poorly described, so it would be worth trying to add some additional tables to make the text more readable. This kind of "result list" does not valorise the Authors' work to the fullest, it's a pity!

7- I had difficulties to understand why some data were collected. I've understood it only once I read the discussion section. I suggest the Authors to insert the reason why they collected, for instance, the vegetation status all around the fallen/snapped tree, before the discussion section.

8 – Here my most difficult request, perhaps. Since the study area is of 0.37 ha only, it would be worthwhile to manage to survey it in its whole. It is important to know exactly the number felled/snapped trees, otherwise it is harder to figure out the true capability of SfM/FMV techniques to properly detect the individual trees. We sure need the ground truths, maybe try to think about doing it in a smaller area but where you will be sure to record every tree affected by the disturbance by using  GNSS tools.

That's all. Thank you!

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

As I myself deal with the problem of forest environment disturbances caused by natural factors and I use material from photogramometric research, I see the problem raised in the manuscript to be significant and useful in practice. The proposed method is promising and the reviewed article is likely to arouse the interest of forest biology specialists and practitioners (e.g. foresters). I assess the research as well designed, and the research area was large and appropriate for the planned study.
The methodology is described clearly and requires no further explanation.
The proposed method has limitations (described in the discussion), but in some respects it exceeds the methods used so far. I especially see the use of FMV in research conducted in natural forests, in the presence of broken and bent trees, with complex vegetation and rich understorey. I suggest that the authors add such possible use in the concluding paragraph of the discussion.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Thanks for your trial to address all the suggestions.

1-   1)  Being the citation of a quite recent review, I would like to see some more effort to explain "Opportunities, Challenges, and Future Directions" in this part of your paper. Please try to add something more.

2-     2) These citations are too old! Any more updates? Other fields of application?

3-    3)  OK. Anyway a question arises: Did you confirm this level of precision by the measurements you did in the field? It was comparable with nominal accuracy of RTK- system? A table with real data and RMSE should be insert here.

4-    4)  Workflow is OK. Figure 1 and 3 are similar to the old ones, really poor. Find other solutions, please. You can't even read words and numbers in Figure 3 !  Some software give the possibility to export the windows, try to do it if possible.

5-     5) OK

6-    6)  OK.  Now it seems more clear and readable

7-     7) OK

8-    8)  OK.  Understandable observations, for clarifying the choices you did.

that's all

Author Response

Please see the attachment

Author Response File: Author Response.docx