Variability in the Spatiotemporal Distribution Patterns of Greater Amberjack in Response to Environmental Factors in the Taiwan Strait Using Remote Sensing Data

Round 1

Reviewer 1 Report

Dear Authors, 

I have read through all of your responses written in the table, and I found that you have addressed all of your concerns and inputs. The responses have been inserted and added to the revised manuscript. Therefore, I have no objection that this manuscript will be published. I am sure this manuscript will be an excellent contribution to fisheries science and will be a piece of good information to the readers. 

Regards, 

Author Response

Thank you for your time and consideration, to give us your valuable suggestions, and reviewed our manuscript.

Reviewer 2 Report

The authors have thoroughly revised the manuscript, and all my previous comments have been addressed. In particular, I appreciate the greater focus on remote sensing in the revised paper. Therefore, I have no further suggestions for improving the paper.

Author Response

Thank you for your effort and time to review our manuscript, thank you for your valuable comments and suggestions to give us to make correct the manuscript.  

Reviewer 3 Report

 

I would like to thank the authors for changing some parts and modifying the figures. The map of bathymetry in figure 1 is very helpful in understanding the discussion. The text is now much easier to follow, and many details were explained. However, my main concern about the considerable difference in the observed and predicted catch rates in winter was not commented on much, not to mention a proper interpretation of that situation. Also, there was no definition of how the year was divided into seasons. Which months were considered spring months? It is particularly important in terms of fish migrations and life cycles. Generally, the seasonal analysis is still very confusing to me which I will explain more below.

According to table 4, “Observed catch rates” were in spring: 372, summer: 378 (almost the same in spring and summer), autumn: 201, and winter 110 (only!). Therefore, based on the in situ observations in winter, the catch rates were the lowest whereas the model predicted the highest ones – even higher than in summer. It was mentioned several times in the Results section and in the Discussion, but there was not even a try to explain that. Lower autumn predictions were decently explained (lines 313 – 317), but not the winter discrepancies. This problem is visible in the three places that I mention below.

 

263 - 267 - “The maps revealed during the summer and winter, indicating the relatively highest predicted catch rates may account for the spatial distribution patterns of S. dumerili in the study area (Figure 8b, d). The catch rate for S. dumerili was predicted to be lower in autumn (Figure 8c). However, catch rates were predicted to increase again in winter in the southern and northern parts of the TS, including in the ECS (Figure 8d).” - when we compare the red dots in winter (8d), higher catches were only in the southern part of TS.

295 - 297 - “This study revealed that the predicted catch rates of S. dumerili distribution were highest in between (23°N–25°N) during spring; the catch rate was high and the area with a high catch rate was widely distributed southward to northward during summer and comparatively lower in autumn, increased catch rates again throughout the winter.”

317 - "However, catch rates were predicted to be higher in the winter, and the majority of fishing locations during this period were in the southern part of the TS, also extending to the northern part of the TS including the ECS."

Then in line 276 I found that in autumn were observed high S. dumerili catch rates, in the northern part of the TS which is completely invisible in Figure 8, and does not agree with the “Observed catches” in Table 4. It might be true for the summer and autumn in 2015 and 2016 (Figure 2), but it can’t be written as a universal statement. I also suggest plotting the dots in Figure 8 in white, and without the outline for clarity:

276 - “High S. dumerili catch rates were observed in the northern part of the TS, extending to 25°N–27°N and 121°E– 124°E, during summer and autumn. Thus, the model-predicted catch rates were reliable, and they correlated well with the observed S. dumerili catch rates in the TS.” – To what data/figure/table exactly does it refer? Why catch rates in Table 4 are the lowest catch rates in the autumn? The red dots in autumn also look bigger in the southern TS than northern. Based on the colour, predicted catch rates in the area were very low, maybe up to 5 kg per hour.

There are also some more issues that I mention below that still need to be addressed in the text:

304 and 307 - "S. dumerili spawn along the edge of the continental shelf in the northern waters of Taiwan in the southern ECS mainly from February to April"; "Furthermore, the high catch rates, particularly during the spawning season, may result in the capture of undersized fish"

"Tone et al. [46] reported similar results, observing a S. dumerili spawning ground in February and March in the ECS."

- What is the division in your study? Please indicate the months considered as spring/summer/winter/autumn (Is winter Dec-Feb? or Nov-Jan?) - It is important in terms of spawning and fish migration in general.

305 - "Our findings suggest that in early spring"? - Having division into four seasons, how can you say it was in EARLY spring?

387 - "Bathymetric surveys conducted in the coastal waters of Taiwan have revealed that the average water depth in the western coastal areas (<50 m) is lower than that in the eastern coastal areas (>2,000 m) [60]. Water depth is a key oceanographic feature that influences the shape and structure of marine species communities, and we observed the highest catch rates for S. dumerili at an MLD of between 20 and 50 m (Figure 4)." - As I understand, the paragraph is about the role of topography and bathymetry, and in the end was mentioned the Mixed layer depth (MLD). In shallow areas, they are somehow related because we can't have MLD deeper than to the bottom, but here the depths in the majority exceeded 50 m. In this case, it looks like terminology and conceptual mismatch. It may be a shortcut, but it confuses the reader. Why is MLD in the same sentence with the bottom depth?

395 - "seasonal monsoons". There was a lot in the text about SST and SSS, but monsoons appear for the first time at the very end. Aren't they related? Why were the monsoons not mentioned before when speaking about significant parameters of the model, such as the SST and SSS?

237-238 - “The interactions between the oceanographic environmental variables in the model revealed that the variables were interdependence on one another” – Shouldn’t it be interdependent?

To summarize, I am surprised that seasonal monsoons are not elaborated more. However, the main concern is discrepancies in the seasonal predictions. Statistics in one thing, but looking at the spatial changes in Figure 8, Figure 2, and Table 4, the red dots do not match in space the high values of predictions, especially in autumn and winter. I work with the models, and I know that there may be some imperfections, but they must be at least carefully explained, much more than they are now.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 4 Report

This manuscript has a high similarity of 25% and the same methodology as the references listed below, which was published in Remote Sensing in 2020:

Naimullah, M., Lan, K. W., Liao, C. H., Hsiao, P. Y., Liang, Y. R., & Chiu, T. C. (2020). Association of environmental factors in the taiwan strait with distributions and habitat characteristics of three swimming crabs. Remote Sensing, 12(14), 2231.

Please find the similarity report and revise the manuscript accordingly before submitting it to the other journal. 

Comments for author File: Comments.pdf

Author Response

Thank you for giving us the opportunity to resubmission of our revised manuscript. According to your comments and reviewers’ comments, for the minor correction, we changed the corrections, point by point, and mentioned in the revised manuscript. We reduced the similarity based on your comments, we checked and reduced it to below 13% similarity. We attached the similarity repot also together.

Round 2

Reviewer 4 Report

Great improvement and can be considered for publication. 

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Dear Authors, 

This is an interesting manuscript. With some little addition and revision, especially on discussion, this manuscript will provide a piece of new information to the readers. 

Regards, 

 

Comments for author File: Comments.docx

Reviewer 2 Report

The manuscript investigates environmental drivers to explain drivers and predict distributions of Seriola dumerili in the Taiwan Strait. I think it is a nice and valuable study, I enjoyed the maps, which are a very good visual tool for conveying information. They could, however,  do with some more work, see detailed comments on figures below. However, I miss a more thorough explanation of the study area – looking at the graphic representations of variables such as SST, there is large variation in temperature in the study area. This could be mentioned in the description of the study area, as well as what the different seasons represent in terms of temperature/monsoons. I would suggest I short section on this in the ‘Methods’ section. I also miss in the Methods section, how many fishing boats this study was based on – if there were any differences in n between the different seasons etc. I find the statistics are appropriate, but some information is missing in the Methods section as pointed out in the detailed comments.   I think the discussion could be organized a bit better for better flow. A few sentences could possibly be moved to the introduction section.

I would also suggest – if the authors would like to go in this direction (as stated in the end of the introduction that the main goal is to improve conservation and management strategies) – to develop these thoughts a bit in the discussion on how findings like this could apply to conservation and management strategies. In some places, the English needs to be revised.

 

 

Abstract

 

Line 22: Maybe “environmental characteristics”?

 

Introduction

 

Line 42: Revise English

 

Line 48: Revise English

 

Line 55: Are these the only existing populations, or just examples of some of the existing populations?

 

Line 60: Could be good to define Kuroshi the first time (a current?), since many readers may not be familiar with this name.

 

Line 62: Remove “,” after “TS

 

Line 65: Add “the” in front of the different current names

 

Line 67: suggest add “the most common”

 

Line 70: if maintain “fisheries” change “this” to “these”

 

Line 79: “address this research gap”

 

Line 80: “a species” or change to “distributions”

 

Line 99: I totally agree that this information could be used for management of this species, but there is no mention in the discussion how this would be useful. It would be nice to see a discussion on this if this is one of the main aims with the paper. (line 99). How can this information be applied to improve management and sustainability of fisheries?

 

Methods

 

Line 105: Remove the “,” after Figure 1

 

Line 106: So the resolution of the fishing data is kg catch per day? If yes, could be good to clarify this.

 

Line 130: Catch percentages? Do you mean catch rates?

 

Line 147: Equation 3: why was latitude and longitude included in the model? Were they included to account for spatial autocorrelation? If yes, could be good to clarify that. Could also be good to justify the choice of GAMs – were all predictor variables non-linear? You could check this with “check.gam” in mgcv, and if not, remove the smoothers for linear predictors.

 

Line 166: Root mean square deviance

 

Figure 2: This figure shows capture per season using data from all years. I am just a bit confused by the colors – what happens if fishing occur in the same location during some of the seasons? E.g. when there is spatial overlap between seasons, which color would be shown in the figure? Does each dot correspond to catch location? Consider also that this combination of colors are not optimal for people with color blindness.

 

Figure 2 caption: 0.1 degree?

 

Figure 3: Looks like some of your variables may be autocorrelated, did you check that?

 

Figure 4: It is nice to see the different variables spatially, but the maps are unfortunately too small for the different catch rates to be visible. You could remove them or you could choose a smaller area as an example to show catch rates maybe. Otherwise there is not really any point to keep them.

 

Figure 5: Methods related to correlation are missing in the methods section. This figure could be put in suppl material, I think.

 

Line 229: The correlation test is missing in the methods section.

 

Line 235: How interaction was tested is missing in the methods section

 

Figure 6: I think this figure could be moved to suppl material. SSH looks kind of linear?

 

Figure 7 caption: The caption is a bit confusing, consider re-write.

 

Figure 8: This is a graphic presentation of the accuracy of the models, which is nice, but it is still difficult to see the sizes of catch rates – which would be one of the points with this figure. I totally understand the difficulty of a large study area, and wonder if perhaps it would be easier to see if you would change the observed values to black instead of white circles? Could be nice with a small picture of your study species as well

 

Figure 8 caption: You could state that “selected GAMs” means p values <0.05. You could also state the RMSE.

 

Discussion

 

Since your findings strongly relates to temperature, climate change would probably have effects on S. dumerili distributions. I understand that this is not the focus of your study, but I think a few sentences on this would further highlight the importance and applicability of your work.

 

Line 282: suggest “..in the TS, where we..”

 

Line 283: A bit unclear what is meant by “evaluating regarding fitting performance..” consider revising this sentence

 

Line 300: what does “higher” refer to?

 

Line 306: This paper seem to refer to squid, any reference on fish?

 

Line 307: Revise English

 

Line 316: But your results suggest that S. dumerili occurs in higher densities in the warmest temperatures? (26-28°C)?

 

Line 324: Could be good to define “high catch rates”

 

Line 326: Consider revise English

 

Line 330: consider define “warm”

 

Line 342: what does the “spatiotemporal results” refer to? Fish distribution?

 

Line 342: Based on what exactly? Would be good to clarify for the reader.

 

Line 350: How can the results from this study be incorporated to help management?

 

Line 352: Distribution patterns based on the catch rates from this study or from other references?

 

Line 361: This part would fit in the introduction or site description to introduce the reader to the scenario

 

Line 383: Revise English. The justification for using Chl a as a predictor would maybe fit better in the introduction or methods section

 

 

Conclusions

 

Line 424: How does this suggestion work with the earlier statement that stocks are overfished? Would increasing the efficiency of the fishing contribute to the primary research goal of the study, which is conservation strategies and fishing management (line 100)?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Reviewer 3 Report

This paper uses oceanographic datasets derived from satellite remote sensing to predict the spatiotemporal distribution of greater amberjack in the Taiwan Strait using generalized additive models. The approach taken is reasonable, the methods are sound, and the paper is mostly easy to follow. My two biggest suggestions for further improving the quality of the paper are 1) to provide some additional context for how the paper contributes to the remote sensing literature; and 2) to do some language editing to correct the mostly minor but numerous grammatical errors throughout the paper. Specific comments are below.

Lines 1-39: In a sign of how the journal fit for this paper is not entirely clear, neither the title, abstract, nor the keywords mention remote sensing at all. Just judging from those indicators, you would expect this paper to appear in a fisheries journal.

Lines 41-100: Along those lines, the introduction provides an overview of the target species and study area, environmental factors influencing fish distributions, and species distribution modeling approaches. It does not discuss how the paper contributes to the remote sensing literature, which should be a prerequisite for a publication in the journal Remote Sensing. At a minimum, the introduction should describe some previous studies that have used remote sensing of oceanographic variables to predict species distributions, explain any gaps in knowledge that still exist, and describe how the current study contributes to filling those gaps. Simply using remotely sensed datasets (all of which have already been processed) is not a sufficient contribution to the remote sensing literature to warrant publication in a remote sensing journal.

Lines 103-105: Where were the fishing vessel data obtained?

Lines 118-119: Here it states that EKE was calculated using the U and V current data, but above it states that EKE was extracted from CMEMS. I assume that the current data was extracted from CMEMS and then the authors used it to calculate EKE? If so, this should be clarified.

Equations 1 and 3: The formatting for these equations seems to be messed up (they are going off the page).

Line 166: Why was 60% of the cumulative frequency chosen as the threshold for high catch rates?

Figure 3: In the y-axis legends for panels a-d, the .0 after the numbers is not needed. That is, panel a can just show 24 and 26 instead of 24.0 and 26.0.

Figure 4: The symbols depicting the catch rates are very difficult to see. Maybe use larger and/or darker symbols?

Table 3: The units for observed and predicted catch rates should be included (I believe they are kg/hour).

Lines 419-420: How is this “consistent with” your findings? It seems that the ranges of EKE found by the two studies were pretty different (0.11 and 0.16 versus 0-0.1).

Lines 433-434: Is it necessarily a good thing that these types of modeling results could be used increase catch? Earlier, it was stated that the species is overfished globally, that its conservation status in the Taiwan Strait is “unclear”, and that there are no management plans for it there. In such circumstances, increasing catch may not be beneficial if it has the potential to increase harvest to unsustainable levels. So it might be helpful to also discuss how information about the environmental controls on habitat may inform any conservation or management plans that may eventually be developed.

Reviewer 4 Report

The article focuses on identifying the main environmental drivers of the greater amberjack (Seriola dumerili) population distribution in the Taiwan Strait. A number of factors were considered, and a generalized additive model (GAMs) was developed for predicting the amberjack distribution to optimize fishing strategy.

Work was explained clearly and in an interesting way. The text needs some language corrections, but the overall impression is good. My main concern is the discrepancy shown in figure 8. Particularly in autumn and winter, catch rates are extremely low compared to the final in situ values. Identification of the factors controlling amberjack distribution was correctly made, and it is an important outcome of the article, but the model and its predictions are either poorly explained or need further tuning.

 

1) Can you please explain the high predictions of catch rates in winter on the Chinese coast (Fig. 8d)?

2) Please explain more the discrepancy between very low estimates of catch rates in autumn and relatively high in situ values.

3) Seasonal differences were analyzed and presented in this article, but the GAMs model was developed for the whole year. Why was that, and have you tested the prediction accuracy for seasonal models?

4) In the text were made several analogies between the greater amberjack and tuna. How are those species similar? How are those references relevant to this work? (lines: 320 – 323; 389 – 391)

 

Some minor comments are mentioned below:

41 - 44 – The first sentence needs to be corrected

equation one - partly invisible

131 - 132 - please add a reference to fig 1 for clarity.

232 - 234 - how would it look like seasonally?

Fig 8 - are the predicted catch rates depicted in colour? Why colour scale ends at 10 when point catch rates exceed 15 kg/hour?- Please add comments in the caption and match the ranges of the scales.

296 - the edge of the continental shelf – in several places are references to the bottom topography. Therefore, please provide this information e.g. by adding an isobaths layer in figure 1 (The black and white frame is sufficient to mark the grid. Instead the grey lattice isobaths will be more useful).

389 - 392 - irrelevant; information in the next sentence (392 - 393) would be sufficient

Reviewer 5 Report

This manuscript has a high similarity of 23% and the same methodology as the references listed below, which was published in Remote Sensing in 2020:

Naimullah, M., Lan, K. W., Liao, C. H., Hsiao, P. Y., Liang, Y. R., & Chiu, T. C. (2020). Association of environmental factors in the taiwan strait with distributions and habitat characteristics of three swimming crabs. Remote Sensing, 12(14), 2231.

Please find the similarity report and revise the manuscript accordingly before submitting it to the other journal. 

Comments for author File: Comments.pdf