Coseismic Gravity Changes and Crustal Deformation Induced by the 2018 Fiji Deep-Focus Earthquake Observed by GRACE and GRACE-FO Satellites

Round 1

Reviewer 1 Report

This paper uses GRACE/GRACE-FO data in terms of spherical harmonic coefficients to obtain the coseismic gravity changes as well as crustal deformations due to 2018 Fiji deep-focus earthquake. The paper is well written, and the topic is important especially considering the lack of GNSS observation network around the study region. The authors have illustrated their idea with proper figures and relevant accurate description. Therefore, publication can be accepted after moderate revision.

1. The abstract requires an improvement. In the present form, the abstract mainly illustrate the motivation of this study, while few of results and the relevant discussions are presented. More detail illustrations and quantified results should be added in current abstract.

2. Line 10: The first was the 2013 Okhotsk earthquake (Mw 8.3) and the second was the 2018 Fiji earthquake (Mw 8.2) on August 19, 2018. 

Line 28: The largest deep-focus earthquake of this century occurred 28 on 19 August 2018 (Mw 8.2) .

Which is exactly the largest deep-focus earthquake?

3. Line 137: suggests – suggest.

4. Line 169: ’The remaining difference was probably caused by the contributions of the 169 second earthquake and the assumption that PREM was used’. The uncertainty of GRACE/GRACE-FO should also be included.

5. The GRACE/GRACE-FO results are only presented in terms of spherical harmonic coefficients, and none temporal signal recovery method is applied after Fan filtering. The leakage in this processing procedure may affect the final conclusion. To avoid this concern, the temporal signal recovery should be considered. Or at least, the comparison should be implemented between spherical harmonic coefficient solution and mascon solution in the supplementary materials.

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The manuscript by Tanaka investigates the coseismic gravity signature of two deep-focus earthquakes occurred in 2018 with magnitudes 8.2 and 7.9 using satellite data from the GRACE and GRACE-FO missions.
The modeling procedure is poorly described and not supported by references showcasing the validity of the approach used here. Furthermore it does not include geometrical and kinematic data about the fault ruptures that are instead retrieved by a series of estimates within tentative ranges of dip angle, rake angle, strike angle and slip amount. I feel that such seismic events have received some scientific attention and perhaps these source parameters have been estimated. Can this modeling approach consider such data rather than tentatively test fault plane geometries? If the answer is no, the author should justify this with a clear and thorough discussion of the modelling approach.
Another major point deserving further attention is the "negligible" gravity contribution provided by the second event (M 7.9) if compared with the first event (M 8.2). The author does not investigate why such impressive difference despite the severe magnitudes of both events.
For all the above I feel that the manuscript needs substantial revision and thus I suggest its rejection.

Here follows a list of detailed comments:

Line 63-75: I find this section a little too poor in details and information required to fully understand and eventually replicate the processing procedure. May be the author can enlarge it and provide some more information to allow the reader not familiar with the procedure to understand it.

Line 83: Was this approach tested by others? Was this formula presented by others or it is the first time that is used? Please expand and strengthen the approach by adding some references showcasing the validity of the procedure used here.
Line 85-91: Please set the correct subscripts for variables and cohefficients.
Also this paragraph is poorly referenced and should be improved to increase readability and clarity for the non-expert readers.

Line 122: Has the author investigated why the second earthquake carries a coseismic gravity change of one order of magnitude lower than the first event? Is this entirely related to the lower magnitude of the second event? Can you discuss this point more in detail?

Line 130: What is this trench? please include it in discussion and figure 1

Line 138: Again, why is it negligible. An event with Mw 7.9 produces a gravity signature of one order of magnitude if compared with an event with Mw 8.2? Why so?

Line 147: Which observations? in lines 138-146 you describe a rectangular fault plane with 100x100 km2 size. Is there any geological or geophysical information supporting this? can you please describe more in detail the "observations" supporting your geometrical model?

Line 155: Again, I feel that such seismic events have received some scientific attention and perhaps the sources have been estimated. Can this modeling approach consider such data rather than assume geometries of the fault plane? If the answer is no, the author should discuss and clarify the modelling approach thoroughly

Line 243: It is not clear to me if the author has tested ALL the fault plane parameters proposed by Fan et al., 2019 and Jia et al., 2020 against the observed gravity signature.

Line 249: This is completely arbitrary and I feel also unsupported by results of this work. You say that the M 7.9 event is negligible on lines 125 and 138. But now you say that the sum of the coseismic gravity changes may be responsible for misfits.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Communication is very interesting, I think it would hold up well as a real article.

The results are very interesting and informative.

I have only one misunderstanding about the text:

line 88: I would appreciate a clarification (in one sentence) of what the Heaviside step function is.

Citations are used in the text using authors' names and year, although the list of citations is numerical.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

This study shows the GRACE/GRACE-FO detection of the coseismic gravity-change signals of the 2018 Fiji earthquake (Mw8.2), which is the second case of successfully detecting the deep-focus coseismic effect by GRACE/GRACE-FO, after the detection of the 2013 Okhotsk earthquake (Mw8.3). The authors obtain clear gravity-increase and gravity-decrease distributions before and after the 2018 Fiji earthquake from GRACE/GRACE-FO observations, and invert a uniform slip rectangle fault slip model according to the observed coseismic gravity changes. This provides new information and insights into the research of gigantic deep-focus earthquakes.

 

The manuscript is well framed, and the methodology is sound. The results are clearly stated. Therefore, I would like to recommend minor revision (mainly related to the addition of some necessary discussions) before publication in Remote Sensing, according to the comments as follows.

 

Comments:

1. The USGS fault slip models appear to significantly underestimate the coseismic gravity changes of the 2018 Fiji deep-focus earthquakes (by comparing Figure 1 and Figure 2). Please consider adding some possible explanations in the Discussion to make the analysis more complete.

 

2. The surface deformation calculation (Figure 4) seems quite independent and not closely relevant to the main conclusions. It might be better to add some descriptions and discussions about the surface deformation results, such as the spatial characteristics and their potential relationship with the fault geometry, or the possibility of validating the calculated displacements with observations at the GNSS sites locating at the surrounding islands (if there are any). Or, considering directly moving Figure 4 into the Supplementary Materials should also be OK.

 

3. The latitude and longitude ranges in the maps of Figure 1 and Figure 2 are not consistent. Please consider using the same latitude and longitude ranges (e.g., 0 to 35S, and 165E to 160W) to show the gravity change distributions in the maps, for the sake of easier comparison of the results.

 

Corrections on writing details:

(1) Lines 17, 20 and 40:

What is the meaning of ‘horizontal distribution’ ? Do you mean ‘spatial distribution’? If yes, please change it into ‘spatial distribution’. If not, please add some descriptions to make it clear.

 

(2) Lines 18 and 59:

GRACE-Follow On -> GRACE Follow-On

 

(3) Line 212:

Figure 3(d)) -> Figure 3(d)

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

I admit that I am a little surprised by the responses of the author mostly because almost nothing changed in the revised manuscript.

If a reader/reviewer tells me that the manuscript I wrote is not fully clear, I do my best to clarify the text and often include also sentences and references that to me may read obvious but perhaps are not so obvious to other readers. In doing so I am motivated by the following: if someone say that the manuscript I wrote is not clear and not convincing, this person may represent a small/large part of the readers potentially interested in my work. Since I'm not writing a paper for myself but for others to understand it, I'll do my best to improve it.
So, if the author feels that comparison with other papers is the best way to write/improve his manuscript, I respectfully disagree because of all the above.

As in the previous revision, I still feel that the manuscript needs adjustments in several parts.

Few examples:
Regarding the formula (1) you wrote that it "is composed of only the terms for overall bias, secular trend, annual and semi-annual seasonal changes, and a step to estimate co-seismic gravity jumps. This linear least square method for time series analysis is widely available and accessible"
the components and the least square approach were clear in the previous version of the manuscript but what was not clear and still remains, is if such approach was used by others to model satellite time-variable gravity signatures.

At lines 87-89 you wrote: "In this equation, the estimated a7 values indicate coseismic gravity jumps at given locations; thus, the spatial distribution of the coseismic gravity change is shown by mapping these values."
In my previous comment about this paragraph I was asking to improve it to "increase readability and clarity for the non-expert readers" and it was motivated also by the aforementioned sentence which is not F=ma but a 7 component least square calculation. Why the a₇ values indicate coseismic gravity jumps? The doubtful reader deserves to find the answer to this question in your manuscript or should he/she go find it somewhere else? if so where?

Line 83: a_i should be written with the subscript i. Same in line 87 for a₇.

"The trench describes the location of the two 2018 Fiji deep-focus earthquakes". I thought that stars were marking the locations of the deep-focus earthquakes... Meaning and description of the "cold front mark" should be stated in caption of figure 1.

I never thought this study was arbitrary. I said that the sentence was arbitrary. Rather than explaining to me the scientific method behind the manuscript, why did the author did not took the chance to include a clear sentence in the manuscript to help the reader understand his reasoning? May be, I guess, he could have included something like ", thus the second earthquake may not be negligible." after he presented the observations in lines 125-132 and maybe, I'm just guessing, recall it in the discussion.

I appreciate the efforts the author made but considering all the above and since little improvements were made in the revised manuscript concerning the points I addressed in the previous version, I suggest major revision.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 3

Reviewer 2 Report

It is interesting that the more this review goes on, the more the author provides useful details that, in my opinion, could be included in the manuscript to increase the value of his work by increasing readability and clarity of its purpose. The unfortunate case is that the author does not include them in the paper but only in the replies to comments, resulting in another review round.

Asking for more clarity of: i) the purpose of your manuscript, ii) of the methods you use and iii) of the relevance of the outcomes, does not sounds to me like “selective reading”, I would rather address it as a “concerned and careful review in the interest of the manuscript”. Furthermore, I never asked to rewrite the manuscript, I only asked to improve clarity and, as a response to my comments, you changed nothing in the first round of review. In the second round of review you made some edits that somehow addressed the relevance of the manuscript, I’m curious about the outcomes of this next round.

In the following I put in italic the text from your replies and I try to separate comments about topics.

 

1-Clarity of the manuscript and of its purpose

I would like to see in the introduction (maybe on line 38) a short paragraph including these information:

this study is the first “clear” detection of coseismic gravity change by deep-focus earthquake… ….The similar studies are, for example, Ogawa and Heki (2007) to detect the post-seismic geoid-height recovery of the 2004 Sumatra-Andaman earthquake (“Slow postseismic recovery of geoid depression formed by the 2004 Sumatra-Andaman Earthquake by mantle water diffusion”, GRL, doi:10.1029/2007GL029340) and Chao and Liau (2019) to detect co- and post-seismic gravity changes by many earthquakes by using least squares and EOF methods (“Gravity Changes Due to Large Earthquakes Detected in GRACE Satellite Data via Empirical Orthogonal Function Analysis”, JGR-Solid Earth, doi:10.1029/2018/JB016862). However, as explained above, they did not detect the clear signal of coseismic gravity change by deep-focus earthquakes.

Following this paragraph I would like also you to say why, in your opinion, these mentioned works do not detect clear signals. In other words, I would like to see why do we need for the approach you are presenting.

I feel that mentioning the above provides some context in which your “novel approach” frames, this, in turn, provides the reader with a baseline to understand it and, as I am doing through this reviews, perhaps appreciate it.

 

2-Clarity and validity of the methods

In my previous comments about the equation (1) I was unsuccessfully trying to ask you to enrich the discussion about the coefficients/constants… My apologies, here I try to rephrase the request: Yes, the least-square approach was clear; yes, the components were stated etc. etc.. But, for example, is there any chance that the least square estimations of a_1-6 constants (coefficients? constants?) provide some error affecting the Δg result? If so, how much does this error affects the standard deviation of Δg? If this is not the case, why? Can you demonstrate it?  

In the previous versions of the manuscript, the formula was provided without any real discussion, almost implying that it is as common as F=m a (see the previous reply to comments), but since yours is a “novel approach”, I would expect some sensitivity analysis to be included/discussed in the main paper. Since this could represent the first “clear detection” of coseismic gravity signals I would like to be sure to review data and not errors. Since this is not the first time I ask for this, I expect some text answering the aforementioned questions with compelling discussion.

Please note that once again, as in the previous comment, this is not a comment to weaken your work, but to try to strengthen it.

 

3-Detailed or specific comments

My previous comment about the comparison with other papers (i.e. the sentence: “So, if the author feels that comparison with other papers is the best way to write/improve his manuscript, I respectfully disagree because of all the above.”) was motivated by the following sentence: “For comparison, in Xu et al. (2021) “Evaluation of GRACE/GRACE Follow-On Time-Variable Gravity Field Models for Earthquake Detection above Mw8.0s in Spectral Domain”, https://doi.org/10.3390/rs13163075 (published by remote sensing) , a less detailed description of similar methods is included”  the point was that I am reviewing your manuscript not Xu’s, so I do not really care about Xu’s work, I care about this work since it is my duty as reviewer.

 

The trench… After two revision rounds this phantomatic trench still remains nameless..

Is there any chance you are referring to the second deepest trench in the world? Is this the Tonga trench? I’m just guessing because, based on the range of latitudes and longitudes of your figures, these should correspond with known location of this trench. So, once you checked it is actually the Tonga trench, why don’t you clearly name it? I feel that, in this case, writing one word, please note, not rewriting the “entire manuscript” but one single word, provides a significant geographical and tectonic reference to your manuscript. Was this a request based on “selective reading”? I do not feel so. Do you need me to precisely mention all the lines where this should be included? OK: 110, 131, 150, 236 and maybe others according to eventual edits. Since in the figures you are showing both the Kermadec (southern part of the cold front you show in figures) and the Tonga (northern part of the cold front you show in figures) trenches, you can mention these with labels in the figures, and/or specifically name both trenches in the text. Alternatively you can label both trenches in the first figure and in text recall only the Tonga trench since it is the closest to your study area. At your choice. Do you need further information about the geological and geographical relevance of the Tonga and Kermadec trenches? Please visit: https://en.wikipedia.org/wiki/Tonga_Trench, https://en.wikipedia.org/wiki/Kermadec_Trench and thousands of scientific papers mentioning them with their proper names.

 

The sentence I referred to as arbitrary was highlighted in the revised manuscript that was attached in the review round 1. Please note that arbitrary does not mean logically incorrect. The sentence I was referring to has been updated in the last round so now it is OK. As per the previous comment about Xu’s work, I do not really care about comments of the other three reviewers. I’m just trying to help in improving the manuscript, the Editors could have ignored my comments and cared only for the other three reviewers, but yet, here we are.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf