PlumeTraP: A New MATLAB-Based Algorithm to Detect and Parametrize Volcanic Plumes from Visible-Wavelength Images
Round 1
Reviewer 1 Report
Dear Editor, Dear Authors
please find below my comments on the revised version of the manuscript with the title
"PlumeTraP: a new MATLAB-based algorithm to detect and parametrize volcanic plumes from visible-wavelength images"
re-submitted by the authors:
Riccardo Simionato, Paul Antony Jarvis, Eduardo Rossi, and Costanza Bonadonna
I have made my comments on the earlier version of the manuscript on the basis they would prove useful for the work carried out as a MSc thesis research project and evolved to a general application. I have revisited the manuscript and I find it a lot more improved and ready for publication ** under the following condition **:
I would recommend the authors to take care of the following points, that are important for the overall merit of the described methodology and quality of the presented work
- lines 135 and on
what is the criterion to select the n value? Is it processing time? compression issues? Please make a clear remark.
- lines 232 and on
The uncertainty is an important point for the validity of the final results. The authors claim that there additional sources of uncertainty due to various factors (human choice, location etc). However, they avoid attempting a quantification. A figure of merit should be provided, even approximately.
- lines 307-308
"... with promising results" For what?
- line 341
(341 +- 5) m to (387 +- 3) made
please use parentheses as the meters unit applies for both numbers, value and error
- lines 430-434
these lines are very similar with 442-444
Please consider merging them.
- lins 511 and on
MATLAB is a licensed based software. If there is a commercial license, please mention it explicitly in the text, if there is a university or lab-wide license, please add an acknowledment.
- Figure A1
I sense a code listing (text) would be a lot better that the screenshot provided. Else, use a larger screenshot. The listing is not easily readable.
Author Response
Response to Reviewer 1 Comments
We thank the Academic Editor and the reviewers for their feedback and comments on the manuscript. We are pleased that the paper has been well-received after our last revision and resubmission. We are also happy to have received constructive reviews that improved the manuscript, and we have made new revisions based on the second round of reviews received.
In this document, we respond to the comments of Reviewer 1, with the review comments in black, and our response in red. Where mentioned, all line numbers refer to the new version of the manuscript which shows the tracked changes. A version without highlighted changes is also uploaded.
We thank Reviewer 1 for their comments and insights throughout that have improved the paper.
Point 1: Lines 135 and on. What is the criterion to select the n value? Is it processing time? Compression issues? Please make a clear remark.Response 1: We tried to clarify in lines 137-138 that the n value (also referred to as frequency factor) is selected by the user at their desired temporal resolution, up to a maximum determined by the frame rate of the original recording. This enables the user to set a sufficient output temporal resolution according to the timescale of the investigated phenomena (there are two examples on lines 143-145). Obviously, as a side effect, the processing time is also determined by the value of the scale factor n. For example, the time needed to analyse a video with n = 2 (i.e., “resampled” at 2 fps) is twice that then for n = 1.
Point 2: Lines 232 and on. The uncertainty is an important point for the validity of the final results. The authors claim that there are additional sources of uncertainty due to various factors (human choice, location etc). However, they avoid attempting a quantification. A figure of merit should be provided, even approximately.
Response 2: Following the reviewer’s comment, we run PlumeTraP with several videos from the available dataset and systematically perturbed the parameters listed in line 236 to quantify the relative variations in the primary outputs, i.e., plume height and maximum width. Therefore, to better explain these sources of uncertainty and their effect on the plume parameters, Appendix B has been added.
Point 3: Lines 307-308. "... with promising results" For what?
Response 3: The “promising results” refer to the capability of PlumeTrap to reproduce an output (a binary image or segmented mask) that reproduces the plume shape. We have added a sentence at lines 310-311 to clarify this point.
Point 4: Line 341. “(341 +- 5) m to (387 +- 3) m”, please use parentheses as the meters unit applies for both numbers, value and error.
Response 4: We thank the reviewer for having noticed this. Following their suggestion, we corrected it (line 345).
Point 5: Lines 430-434 are very similar with 442-444. Please consider merging them.
Response 5: We agree with this comment and have modified lines 442-443 as well as the start of the following paragraph accordingly (lines 445-446).
Point 6: Lines 511 and on. MATLAB is a licensed based software. If there is a commercial license, please mention it explicitly in the text, if there is a university or lab-wide license, please add an acknowledgement.
Response 6: We again thank the reviewer for raising this point. However, we still maintain that there is no need to provide any details on the license agreement universities have with Matlab. Indeed, we have found no publications presenting software tools developed in MATLAB that include such information, (e.g., Pianosi et al., 2015 - 10.1016/j.envsoft.2015.04.009; Valade et al., 2014 - 10.1016/j.cageo.2013.12.015; Bombrun et al., 2018 - 10.1016/j.jvolgeores.2018.01.006). Therefore, we ask the reviewer to share the documentation or references that support their position.
Point 7: Figure A1. I sense a code listing (text) would be a lot better that the screenshot provided. Else, use a larger screenshot. The listing is not easily readable.
Response 7: Figure A1 was resized after the first round of reviews. However, we have now enlarged the figure to how it was before.
PlumeTraP: a new MATLAB-based algorithm to detect and parametrize volcanic plumes from visible-wavelength images by Riccardo Simionato, Paul A. Jarvis, Eduardo Rossi and Costanza Bonadonna.
Reviewer 2 Report
- This a good and well-structured paper. I suggest only modifications on figures. These are very tiny, try to improve theis size and position in the text.
-Send the manuscript to a native English-speaker.
Author Response
Response to Reviewer 2 Comments
We thank the Academic Editor and the reviewers for their feedback and comments on the manuscript. We are pleased that the paper has been well-received after our last revision and resubmission. We are also happy to have received constructive reviews that improved the manuscript, and we have made new revisions based on the second round of reviews received.
In this document, we respond to the comments of Reviewer 2, with the review comments in black, and our response in red. Where mentioned, all line numbers refer to the new version of the manuscript which shows the tracked changes. A version without highlighted changes is also uploaded.
We are pleased Reviewer 2 found our manuscript already in a good shape and structure.
Point 1: I suggest only modifications on figures. These are very tiny, try to improve their size and position in the text.
Response 1: We think the reviewer is referring to the figures of the Appendix A (Figures A1, A3, A4) and maybe Figure 8. We have enlarged these figures to provide a better visualization.
Point 2: Send the manuscript to a native English-speaker.
Response 2: All of the authors of this manuscript have excellent English, whilst one of the authors is a native English speaker. We therefore feel that the quality of the language in this manuscript is more than adequate. Nonetheless, the manuscript has been proof-read again as is good practice.
PlumeTraP: a new MATLAB-based algorithm to detect and parametrize volcanic plumes from visible-wavelength images by Riccardo Simionato, Paul A. Jarvis, Eduardo Rossi and Costanza Bonadonna.
Reviewer 3 Report
I have only a minor question to the authors, which I would like to be addressed in a revised version of the paper:
How do you take into account the direction of the plume in case of winds, when the plume is not a purely vertical structure but tilted in the wind direction. In this case the camera-image plane distance Y is clearly affected in case, when the plume is tilted away from or towards the camera. I assume that the plume height will be clearly underestimated.
One example of such a tilting is e.g. shown in Fig 6c and d - how do you know if the plume is tilted towards or away from the camera? And how do you correct for that?
Author Response
Response to Reviewer 3 Comments
We thank the Academic Editor and the reviewers for their feedback and comments on the manuscript. We are pleased that the paper has been well-received after our last revision and resubmission. We are also happy to have received constructive reviews that improved the manuscript, and we have made new revisions based on the second round of reviews received.
In this document, we respond to the comments of Reviewer 3, with the review comments in black, and our response in red. Where mentioned, all line numbers refer to the new version of the manuscript which shows the tracked changes. A version without highlighted changes is also uploaded.
Point 1: I have only a minor question to the authors, which I would like to be addressed in a revised version of the paper:
How do you take into account the direction of the plume in case of winds, when the plume is not a purely vertical structure but tilted in the wind direction. In this case the camera-image plane distance Y is clearly affected in case, when the plume is tilted away from or towards the camera. I assume that the plume height will be clearly underestimated.
One example of such a tilting is e.g. shown in Fig 6c and d - how do you know if the plume is tilted towards or away from the camera? And how do you correct for that.
Response 1: We thank the reviewer for this pertinent question. This is a factor we considered and have spent considerable time thinking about. As we specified in the manuscript, work to include the effect of wind direction in PlumeTraP is ongoing (line 433). However, in this manuscript, we wanted to focus on the image processing routine and so this version of PlumeTraP is released without the wind correction. The wind correction itself will be based on a tool developed by a different author that has just been submitted (Snee et al. - submitted to Volcanica). Therefore, the results obtained with this version of PlumeTraP should be taken carefully (lines 431-432). Once the Snee et al. manuscript is published, we plan to release a new version of PlumeTraP including the wind correction tool, along with a detailed user’s manual. We have also stated this in the Data Availability Statement section of the manuscript (line 514) and have added a brief description of how wind affects the measurements (lines 426-429).
PlumeTraP: a new MATLAB-based algorithm to detect and parametrize volcanic plumes from visible-wavelength images by Riccardo Simionato, Paul A. Jarvis, Eduardo Rossi and Costanza Bonadonna.
Author Response File: 
Author Response.pdf
