Optimizing Source Wavelets Extracted from the Chirp Sub-Bottom Profiler Using an Adaptive Filter with Machine Learning

Round 1

Reviewer 1 Report

Dear authors,

I read the paper with great interest, the implementation of ML to extract the wavelet of chirp SBP data, which has a potential for relevant processing improvements in HR seismic. However, the paper in the present form is not suitable for publication, and I have recommended major revision. The text needs reorganization and focus, as I suggested in my comments (attached file).

Many parts of the text are not in their proper section. Many parts of the text can be significantly reduced. The abstract highlights the chirp data processing, but needs better delineations of the study with some elements of objectives and methods that must be presented. Introduction has many repetitions of irrelevant contents. The initial rationale goes further out of the subject and is highly general. This section must focus on why extract the wavelet of specifically chirp data is important. The content of materials and methods are note clearly grouped into one unique section (but highly spread), that difficult distinguish results from methods, figures of seismic sections must be improved specially where changes in processing are shown (frames with more zoom can help). Description of the chirp SBP used on survey and the bottom/sub-bottom types imaged should be present. Computational processing in Matlab must be better presented, and a supplementary material with routines would be interesting for reproduction. I suggested Material and Methods instead “basic theory”, see the pdf comments. Result. This section can also be rewritten, providing only the description of the implementation and analysis of the products of the processing. These products need to be discussed in terms of methodological approaches, including other seismic sources, and the visual quality/improvements of the seismic sections. The authors should consider separate the discussion and results. 

Comments for author File: Comments.pdf

Author Response

Response to the Reviewer’s Comments

Reviewer #1_Round 1

 

Reviewer(s)' Comments to Author:

 

Dear authors,

 

I read the paper with great interest, the implementation of ML to extract the wavelet of chirp SBP data, which has a potential for relevant processing improvements in HR seismic. However, the paper in the present form is not suitable for publication, and I have recommended major revision. The text needs reorganization and focus, as I suggested in my comments (attached file).

 

Many parts of the text are not in their proper section. Many parts of the text can be significantly reduced. The abstract highlights the chirp data processing, but needs better delineations of the study with some elements of objectives and methods that must be presented. Introduction has many repetitions of irrelevant contents. The initial rationale goes further out of the subject and is highly general. This section must focus on why extract the wavelet of specifically chirp data is important. The content of materials and methods are note clearly grouped into one unique section (but highly spread), that difficult distinguish results from methods, figures of seismic sections must be improved specially where changes in processing are shown (frames with more zoom can help). Description of the chirp SBP used on survey and the bottom/sub-bottom types imaged should be present. Computational processing in Matlab must be better presented, and a supplementary material with routines would be interesting for reproduction. I suggested Material and Methods instead “basic theory”, see the pdf comments. Result. This section can also be rewritten, providing only the description of the implementation and analysis of the products of the processing. These products need to be discussed in terms of methodological approaches, including other seismic sources, and the visual quality/improvements of the seismic sections. The authors should consider separate the discussion and results.

 

Reply:

Thank you for your detailed and insightful review of our manuscript. We have reflected deeply over your comments and made the necessary revisions to the manuscript. Our detailed responses to your comments are as follows:

 

 

 

Reviewer’s Comments and responses for the "Abstract" part.

 

1. line 8-9

“Marine seismic profiling with a Chirp sub-bottom profiler (SBP) is primarily used for sub-bottom stratigraphic surveying owing to its high-resolution and low penetration depth.”

It seems that this introductory sentence is irrelevant. The paper focus on processing with specific algorithims and need an appropriated context of the problem.I suggest a breef "technical" view of the chirp data. It is an enveloped data? uncorrelated? minimum phase wavelet? what is the challenge to wavelet extraction? So, point to the specific problem/context that deserve your study.

 

Reply:

Thank you for your comment. Following your suggestion, we have revised the Abstract.

 

2. Line 10

“we developed a novel approach for applying seismic data processing techniques to raw data.”

Again, it is not a satisfatory affirmation ("processing to raw data"). This objective is vage.

 

Reply:

Thank you for your comment. Following your suggestion, we have completely revised the Abstract.

 

3. Line 11

“processing techniques”

avoid repetition

 

Reply:

Thank you for your comment. Following your suggestion, we have removed repetition.

 

4. Line 13

how was done? which software? or programing language?

 

Reply:

Thank you for your comment. We used Matlab for our data processing. This information has been added to the text.

 

5. Line 14

raw?

 

Reply:

Thank you for your comment. Yes, we meant ‘raw data’.

 

6. Line 15

“two different adaptive filters.”

Specify

 

Reply:

Thank you for your comment. Following your suggestion, we have detailed this part.

 

7. Line 17

“applied with an adaptive filter to the primary reflection signal using 17 the ideal source wavelet as a reference waveform.”

I can imaging that the machine learng was applyed at this point of the flow processing. So, you should give a little descriprion of the function/algorithm that maps/classify the wavelet inputs to outputs. this is the main part of your inovation, and needs more enphasis

 

Reply:

Thank you for your comment. Following your suggestion, we have described the data processing process in detail.

 

8. Line 20

“novel approach”

avoid repetition

 

Reply:

Thank you for your comment. Following your suggestion, we have revised this to avoid repetition.

 

9. Line 21

“problem of source wavelet extraction.”

first clear assing this issue at the begining. The abstract is not clear in general. It needs some improvements, clear show the problem addressed (focused on chirp data), state the objectives, give better descriptions of what and how was done, particularly the machine learnig training and application. The final statement should point futher researches (not back to the objectives)

 

Reply:

Thank you for your comment. Following your suggestion, we have completely revised the Abstract.

 

Reviewer’s Comments and responses for the "1. Introduction" part.

 

10. Line 25, 31, 38

“Marine seismic profiling”

repetition... 1st, 2nd, 3rd paragraphs start as the same.the 1st paragraph is really not relevant, could reduced to 1 sentecne and jointed to the 2nd paragraph.I would suggest describe the chirp pulse, the envelope matters, processing...

 

Reply:

Thank you for your comment. Following your suggestion, repeated phrases have been removed.

 

11. Line 39

“quaternary”

Q

always uppercase

 

Reply:

Thank you for your comment. We have capitalized “Quaternary.”

 

12. Line 40

“archaeology.”

Again, this first sentence is introducting repeatedly.

 

Reply:

Thank you for your comment. The text has been modified to avoid repetition.

 

13. Line 44-45

“about sea bed and seafloor”

choose one term only

 

Reply:

Thank you for your comment. Following your suggestion, we have selected just one term.

 

14. Line 46-47

“As such, many studies on seabed classification and seabed sediment character-46 ization have been conducted using data acquired via Chirp SBPs [9,12,13].”

It is not properly correct. The classification of bottom sediments is mainly by MBS and SSS. sub-bottom information is not necessary.Otherwise, up to this point of the intro, the text did not developed an understanding of why extract the wavelet of chirp data is important.I suggest rewrite the entire introduction into a discussion about the relevance of the original wavelet extraction from enveloped data, and then its use in the processing.Besides,  would be interesting argue why the processing of this type of data is important (once we can use it without processing).

 

Reply:

Thank you for your comment. Following your suggestion, we have revised this unnecessary sentence.

 

14. Line 52

this paragraph is disconected with the previous one. and also not directly relevant

 

Reply:

Thank you for your comment. Following your suggestion, we have deleted this text.

 

15. Line 55

“resolution improvement through computational processing,”

1. the paragraph (and the above) should start from this issue: the computational processing applied to specific improvements on seismic data.

 

Reply:

Thank you for your comment.

 

16. Line 62

great!

This is the subject. I strongly suggest to develop the Introduction

since the begining of the text discussing the meaning, relevance, needs, and possibilities.

much of the precedding paragraphs are pointless to you real topic and could be drasticaly reduced.

 

Reply:

Thank you for your comment.

 

17. Line 64

“Klauder wavelet”

give quick explanation

 

Reply:

Thank you for your comment. Following your suggestion, we have rearranged the text to provide an explanation here.

 

18. Line 67

need citation

 

Reply:

Thank you for your comment. Following your suggestion, we have added a citation.

 

19. Line 72

“adaptive filter”

based on?

 

Reply:

Thank you for your comment. Following your suggestion, we have revised the text to add more detail.

 

20. Line 74

“applied”

use the terms of ML.

is there any statistical method ?

make clear

 

Reply:

Thank you for your comment. Following your suggestion, we have revised the text to enhance clarity.

 

21. Line 85

put a conclude remark here.

 

Reply:

Thank you for your comment. Following your suggestion, we have revised this text.

 

Reviewer’s Comments and responses for the "2. Basic theory" part.

 

22. Line 86

here I suggest Material and Methods instead "basic theory"we need a brief description of the area surveyed (which material is the bottom? muddy, sandy? rocky?); the equipement specification and parameters from the survey; them what/how was done the processing with ML.About the processing and the use of ML, what was the software for aquisition? and for processing? and to implentation of the ML? or code language? all of these things should be well described in this section of Materials and Methods. I suggest to include as suplementary material the code exampleMust there is enough information for replication

 

Reply:

Thank you for your comment. Following your suggestion, these sentences have been revised.

 

23. Line 89-97

“The first Chirp SBP was developed by LeBlanc and Mayer for the quantitative classification of seafloor sediments [6, 9]. In general, Chirp SBPs can be used to obtain relatively detailed geological structural information of sedimentary environments, such as determination of the marine sediment thickness, identification of seafloor faults, or sub-bottom structure studies in shallow and coastal waters. Additionally, the technique has been used to identify risk zones for the installation of offshore structures, such as in surveys of gas leaks from the seabed. Chirp SBPs are useful for finding ideal locations for structures such as subsea pipelines, offshore plants, and wind farms and implementing their designs [10,12,16, 18].”

You start the paragraph with "high-resolution sub-bottom profilers that use wide-band frequency modulation signals as a sound source" So, we expect you explain the chirp signal. all of history and aplications you already wrote in Introduction topic. remove from here.

 

Reply:

Thank you for your comment. Following your suggestion, these sentences have been revised.

 

24. Line 105

“Figure 1”

not necessary. Remove

 

Reply:

Thank you for your comment. Figure 1 has been removed.

 

25. Line 127

First indicate the equation then the elements

 

Reply:

Thank you for your comment. Following your suggestion, this text has been revised.

 

26. Line 128-130

use italic or ""

 

Reply:

Thank you for your comment. Following your suggestion, this text has been revised.

 

27. Line 149

symbology need to be checked according to the journal rules

 

Reply:

Thank you for your comment. Following your suggestion, this text has been revised.

 

Reviewer’s Comments and responses for the "3. Results and discussion" part.

 

28. Line 188

“2 to 7 kHz”

FM pulse?

which is the manufacture?

towed, with pressure sensor, pulse rate, sampling?...

 

Reply:

Thank you for your comment. Following your suggestion, we have added details regarding the equipment used in this study.

 

29. Line 196

“2 ms.”

up to here you present the "Material", not Result. do not represente what was done.this part should be moved to Materials and Methods.

 

Reply:

Thank you for your comment. Following your suggestion, this text has been

moved to Materials and Methods.

 

 

30. Line 204-205

“which was set with a reflectivity coefficient of +1 at 0.018 s and -0.5 at 0.027 s.”

Now you jumped to parametrization, but we dont know about the survey, the equipement, the geological context where the seismic data was aquired.which stratigraphy? the preceding text do not allow understand as a result.

 

Reply:

Thank you for your comment. Following your suggestion, this text has been revised.

 

31. Line 214

“Quoted from the SEG Terminology Dictionary”

All conceptual must be in the previous topic. Here we expect the new way ML and its products

 

Reply:

Thank you for your comment. Following your suggestion, this text has been revised and moved to the Induction.

 

32. Line 229

“Figures 5–7 show”

However you did not present any description of these figures. the descriptions must be here, before the figures appear.

 

Reply:

Thank you for your comment. Following your suggestion, this text has been

modified to add additional explanations and to make it easier to understand.

 

33. Line 230

“SeisSpace ProMAX-Landmark Solutions program using”

present this in the topic of mat./methods

 

Reply:

Thank you for your comment. Following your suggestion, this text has been revised, and we have removed the unnecessary part.

 

34. Line 245-249

“the results be used as basic analysis data in stratigraphic classification or bed-rock depth studies, for which Chirp SBPs have been widely used. However, if the source wavelet for deconvolution processing is extracted using raw data, an improved section can be obtained. Such processing produces results suitable for profiling purposes, such as the detection of buried objects or stratigraphic classification.”

keep on data signal characteristics and processing limitations.

 

Reply:

Thank you for your comment. Following your suggestion, this text has been revised.

 

35. Line 251

“Raw data is necessary to process Chirp SBP data.”

please clarify or rewrite. Obviously we process the raw data, but you mean the source wavelet?.

 

Reply:

Thank you for your comment. Following your suggestion, this text has been revised.

 

36. Line 259-264

“For such source wavelet recording and analysis technology, professional profiling institutions in some countries, such as Bolt in the USA, WesternGeco in Norway, and the Geological Survey in Canada, hold source wavelets for thorough quality control of marine seismic profiling. Currently, the Korea Institute of Geoscience and Mineral Resources is also trying to secure proprietary technology by measuring and analyzing air gun source wavelets [24].

this is the content that I would move to the introduction. It is useful to argue the relevance of you topic.However it is not a result, or a discussion about the meaning of your results.

 

Reply:

Thank you for your comment. Following your suggestion, this text has been revised to make it easier to understand. It has also been moved to the Introduction.

 

37. Line 273

“using a bubble pulsar”

It is not a mixed-phase wavelet? (it is similar to the sparker

if yes, it is not comparable to the chirp. It deserves better discussion

 

Reply:

Thank you for your comment. Following your suggestion, this text has been revised and the unnecessary parts have been removed.

 

38. Line 276

“we used the MATLAB-based GUI tool to extract”

move to the method section

the routine could be in suplementary materials

 

Reply:

Thank you for your comment. Following your suggestion, this text has been revised.

 

39. Line 278

“at a center frequency of 4.5 kHz”

how it was defined? frequency analysis? Describe

 

Reply:

Thank you for your comment. Following your suggestion, this text has been revised to include a more detailed explanation.

 

40. Line 337

“①, ②, and ③ show”

for instance, use 'a' in Fig 13

Need much more zoom to see the changes you are pointing.

 

Reply:

Thank you for your comment. Following your suggestion, we have revised the figures.

 

41. Line 347

need better zoom for visual avaliation and comparizons (observable changes in the sesimc section are the pratical result we expect). The figures are in to large scale and almost identical. give more zoom.

but I would also suggest an statistical approach...

 

Reply:

Thank you for your comment. Following your suggestion, we have revised the figures. We agree that taking a statistical approach would be useful; this will be a focus of our future work.

 

42. Line 385

“such as matching filters, not raw data.”

must be previously explained. see other comments

 

Reply:

Thank you for your comment. Following your suggestion, we have revised this text to make it easier to understand, and unnecessary text has been removed.

Author Response File: Author Response.pdf

Reviewer 2 Report

• The abstract is mainly focused on the methodology part; however, the key results of the work are not properly presented.
• Please add some references in the first paragraph of the introduction.
• The introduction should contain at lease some information about the application of machine learning in similar studies.
• The authors should add a separate chapter of the methodology. I suggest to merge it with the chapter of Basic theory.
• Line 188: The frequency band of the sound source used in this study was 2 to 7 kHz, and the pulse length was 2 ms. Please explain why the authors have selected these parameters.

Author Response

Response to the Reviewer’s Comments

Reviewer #2_Round 1

 

Reviewer(s)' Comments to Author:

1. The abstract is mainly focused on the methodology part; however, the key results of the work are not properly presented.

Reply:

Thank you for your comment. According to your suggestion, we have completely revised the Abstract.

 

2. Please add some references in the first paragraph of the introduction.

Reply:

Thank you for your comment. According to your suggestion, we have revised the Introduction.

 

3. The introduction should contain at lease some information about the application of machine learning in similar studies.

Reply:

Thank you for your comment. According to your suggestion, we have adding additional explanations in the Introduction.

 

4. The authors should add a separate chapter of the methodology. I suggest to merge it with the chapter of Basic theory.

Reply:

Thank you for your comment. The text has been modified so that this section is included within the Materials and Methods section.

 

5. Line 188: The frequency band of the sound source used in this study was 2 to 7 kHz, and the pulse length was 2 ms. Please explain why the authors have selected these parameters.

Reply:

Thank you for your comment. According to your suggestion, this text has been revised to include more detailed discussions and explanations.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Dear authors, 

The manuscript is now better presented. Many repetitions and irrelevant contents were removed. There are some minor revisions need. An English review could improve the text overall. After the new revision, I would recommend for publication.

 

Comments

line 19: “...the approach developed in this study addresses issues associated with extracting source wavelets”. This last abstract line is a repetition of your objective. I would suggest to provide the direct implication or general conclusion.

line 27-28: these two terms are still weird and do not encompass what the text requires. “sedimentation” is related with sediment dynamics and modern processes, while “structural” relates to faults, folds, hazards... that is different from earth structure... of course seismic is not limited to these applications. Please, rewrite and adjust like “The high-resolution reflection seismic technique is widely used to onshore and offshore site investigations...” or you can say “The high-resolution reflection seismic technique is widely used from shallow to deep-water marine site investigations...”. This is a keyword that didn’t appeared yet “high-resolution reflection seismic”

line 31: Archaea? I don’t see any seismic being used to surveys of single-celled organisms. Marine geology, geotechnical, geohazards, sediment dynamics, petroleum industry, environmental research etc... are better applications for shallow seismic. In this list, you could include the archaeological surveys.

line 34: change “fingers” to “pingers”

line 35-36: you already gave examples of marine seismic in the line 30-31. Here, “Quaternary studies” is enough. Connect to the next sentence.

line 36: connect to the end of the sentence in lines 32-34. Rewrite like this: “Besides, the chirp enveloped data allows sub-bottom interpretations without additional processing being successfully applied to Quaternary studies.”

line 37: delete the entire sentence.

line 39: change “submarine” to “seafloor classification”

Lines 38-43: Must be improved. Rewrite completely or delete. The only information relevant is that the SBP can support some acoustic seabed substrate classification.

Lines 44-47: the paragraph is incomplete. Join to the next.

line 48-52: reduce and move to the part of examples from line 31.

line 169: remove the repetition. Repeated from line 167.

line 180: Don’t you used the SeisSpace ProMAX-Landmark Solutions program? give the description.

line 183: remove. It is clear from the previous paragraph.

lines 190-191: this is repeated three times. Consider keep this sentence in only one place.

line 196: it is not “stratigraphic”, that is geochronological time and geometry of layers. This acoustic, hydroacoustic, or seismic, or even seismo-stratigraphic that is geophysics. Choose a better geophysical term.

line 234: a little confuse. “acoustic wave signature”? make clear.

line 259: imprecise language.

Comments for author File: Comments.pdf

Author Response

The manuscript is now better presented. Many repetitions and irrelevant contents were removed. There are some minor revisions need. An English review could improve the text overall. After the new revision, I would recommend for publication.

 

Reply:

Thank you for your detailed and insightful review of our manuscript. Thanks to your review, we were able to substantially improve the manuscript.

We are very grateful.

 

Comments

 

line 19: “...the approach developed in this study addresses issues associated with extracting source wavelets”. This last abstract line is a repetition of your objective. I would suggest to provide the direct implication or general conclusion.

 

Reply:

Thank you for your comment.

The sentence was modified as follows:

“Our results highlight the effectiveness of performing deterministic deconvolution by extracting source wavelets using adaptive filters, and we believe that our method is useful for marine seismic exploration data processing.”

 

line 27-28: these two terms are still weird and do not encompass what the text requires. “sedimentation” is related with sediment dynamics and modern processes, while “structural” relates to faults, folds, hazards... that is different from earth structure... of course seismic is not limited to these applications. Please, rewrite and adjust like “The high-resolution reflection seismic technique is widely used to onshore and offshore site investigations...” or you can say “The high-resolution reflection seismic technique is widely used from shallow to deep-water marine site investigations...”. This is a keyword that didn’t appeared yet “high-resolution reflection seismic”

 

Reply:

Thank you for your comment.

We revised the sentence by reflecting your review.

 

line 31: Archaea? I don’t see any seismic being used to surveys of single-celled organisms. Marine geology, geotechnical, geohazards, sediment dynamics, petroleum industry, environmental research etc... are better applications for shallow seismic. In this list, you could include the archaeological surveys.

 

Reply:

Thank you for highlighting this typo.

The term was incorrect.

We wanted to refer to underwater archaeology.

 

line 34: change “fingers” to “pingers”

 

Reply:

Thank you for your comment.

We corrected the word as suggested.

 

line 35-36: you already gave examples of marine seismic in the line 30-31. Here, “Quaternary studies” is enough. Connect to the next sentence.

 

Reply:

Thank you for your comment.

As suggested, we revised the respective sentence for conciseness and clarity.

 

line 36: connect to the end of the sentence in lines 32-34. Rewrite like this: “Besides, the chirp enveloped data allows sub-bottom interpretations without additional processing being successfully applied to Quaternary studies.”

 

Reply:

Thank you for your comment.

We revised the sentence for clarity.

 

line 37: delete the entire sentence.

 

Reply:

Thank you for your comment.

We deleted the sentence.

 

line 39: change “submarine” to “seafloor classification”

 

Reply:

Thank you for your comment.

We revised the term as suggested.

 

Lines 38-43: Must be improved. Rewrite completely or delete. The only information relevant is that the SBP can support some acoustic seabed substrate classification.

 

Reply:

Thank you for your detailed review of your manuscript.

As suggested, the lines referred to were repetitive, so that we deleted them.

 

Lines 44-47: the paragraph is incomplete. Join to the next.

 

Reply:

Thank you for your comment.

We revised the paragraph and move it up for clarity.

 

line 48-52: reduce and move to the part of examples from line 31.

 

Reply:

Thank you for your comment.

This part was moved as suggested.

 

line 169: remove the repetition. Repeated from line 167.

 

Reply:

Thank you for your comment.

The repeated content was deleted, and “2 ms of pulse length” was added at the beginning of the sentence to avoid repetition.

 

line 180: Don’t you used the SeisSpace ProMAX-Landmark Solutions program? give the description.

 

Reply:

Thank you for your comment.

We did not use the Promax software.

We used an own program developed with the Matlab G.U.I.

We described in detail in "Abstract" and "2.2. Acoustic Characteristics of a Chirp Wavelet".

 

line 183: remove. It is clear from the previous paragraph.

 

Reply:

Thank you for your comment.

As suggested, the respective passage was deleted for clarity.

 

lines 190-191: this is repeated three times. Consider keep this sentence in only one place.

 

Reply:

Thank you very much for your careful review.

We deleted any repetitions in the respective passage.

 

line 196: it is not “stratigraphic”, that is geochronological time and geometry of layers. This acoustic, hydroacoustic, or seismic, or even seismo-stratigraphic that is geophysics. Choose a better geophysical term.

 

Reply:

Thank you very much for your careful review.

We agree with your comment. Therefore, we have revised the text from "The Stratigraphic Information" to "The information on the bottom layer of the seafloor".

 

 

line 234: a little confuse. “acoustic wave signature”? make clear.

 

Reply:

Thank you very much for your careful review.

We sincerely apologize for the confusion and the mistake.

It has to be “seismic expression”.

Moreover, we revised "of" to "from" at the beginning of the sentence.

 

line 259: imprecise language.

 

Reply:

Thank you for your comment.

We want to refer to "strata sections" rather than to "deep stratographic sections". As a result of a careful review according to your comment, we revised the text accordingly.

Author Response File: Author Response.pdf