Spatio-Temporal Variability of the Habitat Suitability Index for the Todarodes pacificus (Japanese Common Squid) around South Korea
Round 1
Reviewer 1 Report
General comments:
In this manuscript performed the habitat simulation by using commercial catch data and satellite datasets for the Janpanese Common Squid. This manuscript provides some good information on the fishing grounds with high catch probability and long-term variations in spatial and temporal adaptability. This is an interesting issue for scientists and aquatic managers. The manuscript is well-written and the idea is clear. However, the organizing the paragraph is awkward. Thus I recommend the manuscript “Major Revision.” In that sense, comparisons with existing results of relevant this manuscript may be more informatics and persuasive of the author’s findings. Some possible directions for improvement are listed below.
Specific comments:
(1) This also seems a very local study, with no implication elsewhere: to increase the scope of this manuscript, and not being once again a local case-study, it would be useful to point out here, why this study is important, i.e. how the findings of the present study can be applied elsewhere. This should be included in the last sentence of the abstract as well as in the Discussion and Conclusion. What is indeed the novelty of this study? Can this be applied elsewhere? You should focus on these messages to the readers, as it is what they want to really know.
(2) Please formulate clear objective, and further interest of this work. Is it intended to create a protocol that can be useful in many other projects of water quality, and habitat change of the aquatic species?
(3) It is a good idea to state what impacts you have on other fish species and whether you have any future plans.
(4) In the methods there are lacks of descriptions of the target species characteristics, sample sites, methods and techniques for sampling, how, when, where etc. this target species was sampled. Perhaps it would be good the authors providing some information about ecology of the target species (i.e. type of habitat, reproduction timings, tolerance to human pressures, etc.).
(5) It is very natural that product equation can have lower habitat area compared to geometric men, because geometric mean considers compensation between high and low suitability values. Although the reviewer understands this situation as one stage of the research, it should be mentioned somewhere in the manuscript.
(6) Conclusions: in the present form this chapter is rather a summary than conclusions, describing the study processes instead of giving ‘take home messages’ from the results of the study for the reader.
Author Response
Specific comments:
(1) This also seems a very local study, with no implication elsewhere: to increase the scope of this manuscript, and not being once again a local case-study, it would be useful to point out here, why this study is important, i.e. how the findings of the present study can be applied elsewhere. This should be included in the last sentence of the abstract as well as in the Discussion and Conclusion. What is indeed the novelty of this study? Can this be applied elsewhere? You should focus on these messages to the readers, as it is what they want to really know.
- In this study, we derived the habitat suitability index for the T. pacificus based on the species’ preferred environmental conditions. Also, this model could be applied to global ocean. In that case, several suitable habitats will be found in many other regions. However, T. pacificus distributed mostly around South Korea and Japan (or northwestern pacific). Thus, the global application of the model in this study can be misleading to readers. Therefore, this approach would be applied elsewhere for other target species after a modification with in-situ fishery dataset. Also, we think the advantage of this study is the simplicity of the algorithm. Furthermore, there are only a few studies that have used primary production as an environmental variable for the HSI model. The sentences were revised (Line 29-31, 73-75, 228-234).
(2) Please formulate clear objective, and further interest of this work. Is it intended to create a protocol that can be useful in many other projects of water quality, and habitat change of the aquatic species?
- Our further interest is to examine long-term variations of the HSI. As mentioned in Introduction, the distribution of fishery resources around South Korea is rapidly changing. Moreover, some species are disappeared in Korean waters which are important food for Korean. Thus, our ultimate goals will be the long-term analysis of the HSI for target species. Our further interest and goals are added in section 5. Summary and conclusion (Line 292-295).
(3) It is a good idea to state what impacts you have on other fish species and whether you have any future plans.
- Our further interesting and ultimate goals are added in summary and conclusion (Line 292-295).
(4) In the methods there are lacks of descriptions of the target species characteristics, sample sites, methods and techniques for sampling, how, when, where etc. this target species was sampled. Perhaps it would be good the authors providing some information about ecology of the target species (i.e. type of habitat, reproduction timings, tolerance to human pressures, etc.).
- The descriptions for the target species were added in introduction (Line 41-51)
(5) It is very natural that product equation can have lower habitat area compared to geometric men, because geometric mean considers compensation between high and low suitability values. Although the reviewer understands this situation as one stage of the research, it should be mentioned somewhere in the manuscript.
- As mentioned in method, we used arithmetic mean model because many several studies have reported that the AMM is the most appropriate model for the HSI. The product equation (eq 4) was incorrectly written. We revised equation 4.
(6) Conclusions: in the present form this chapter is rather a summary than conclusions, describing the study processes instead of giving ‘take home messages’ from the results of the study for the reader.
- We revised the paragraph (Line 292-295).
Reviewer 2 Report
[General comment]
The manuscript by Lee et al. developed the model for estimating the habitat suitability index (HSI) of Todarodes pacificus (Japanese Common Squid) around South Korea based on satellite-derived variables. The authors investigated the seasonal habitat patterns of T. pacificus using the developed HSI model and showed the hotspots in their study region. However, although the manuscript is an important study for sustainable fisheries management, the authors should clearly show new insights and findings that have not been provided in past studies. The justification for this manuscript is unclear in the current form. I have the following concerns.
[Major comments]
This manuscript and Lee et al. (2018, remote sensing) are basically the same except for the target species. What is the main difference? Although the authors developed the HSI model for pacificus with four satellite-derived variables, is it better at capturing hotspots compared with the previous studies, e.g., Arabia et al. (2016, remote sensing)? Is it statistically more accurate? These points should be clearly described and stated in the manuscript. As it stands, it is hard to understand why this study is needed and to follow what is the essential information and new insights described in this study.
The hotspots with more than 0.7 of HSI (Figures 7 and 10) in August and December are not consistent with high values of total fishery ladings and the number of fishing records in August and December described in Figure 2. What is the main factor affecting the gaps between the landings and the model output?
VGPM depends on the variables of SST and chlorophyll a concentration, which are overlapped with the parameters used in the HSI model. Furthermore, in this study, SST values were classified into two segments, winter–spring and summer–autumn. The explanation for the influence of this partition on the results of PP values needs to be more explicit.
The authors employed 8-day composite images of the Level 3 MODIS data as described in Lines 77–79 of the method section. How did the authors calculate monthly composite images and the HSI values represented in Figure 7? This point should be clearly stated in the manuscript.
[Minor Comments]
Line 22: SST and SSHA
SST and SSHA should be spelled out when the term appears in the text for the first time.
Lines 111 – 115: Although 37 % of the catch data could not…
This sentence is simply repetition and should be deleted.
Lines 118, 125, 171,173: (Figure 3), (Figure 4) …
Please check the figure number through the overall manuscript and revise it accordingly.
Figure 10
Figure 10 is redundant as this information already appears in Figure 7.
Author Response
This manuscript and Lee et al. (2018, remote sensing) are basically the same except for the target species. What is the main difference? Although the authors developed the HSI model for pacificus with four satellite-derived variables, is it better at capturing hotspots compared with the previous studies, e.g., Arabia et al. (2016, remote sensing)? Is it statistically more accurate? These points should be clearly described and stated in the manuscript. As it stands, it is hard to understand why this study is needed and to follow what is the essential information and new insights described in this study.
This manuscript and Lee et al. (2018) are deriving the HSI for the target species with same way. The main difference is environmental parameters. In this study, SSHA was considered as an additional environmental parameter. Moreover, due to the migration characteristics of T. pacificus, suitability indices of SST and SSHA for Winter-Spring and Summer-Autumn groups are derived separately. It is hard to determine which way is more accurate than other previous studies , but we think the advantage of this study is the simplicity of the algorithm. Furthermore, there are only a few studies that have used primary production as an environmental variable for the HSI model. The sentences were added in lines 73-75, 228-231.
The hotspots with more than 0.7 of HSI (Figures 7 and 10) in August and December are not consistent with high values of total fishery ladings and the number of fishing records in August and December described in Figure 2. What is the main factor affecting the gaps between the landings and the model output?
The unordinary high catch was recorded in August 2015. Without the record, total fishery landings in August is not much high. The record was excluded in the process of algorithm derivation because it is outlier. However, it was included in Figure 2. We revised the figure to avoid misunderstandings. Most of fishing records for December were reported in late December, and were found in the southern part of the East/Japan Sea and the western part of the South Sea. Actually, the HSI begins to increase in late December. However, our figure describing the climatological monthly mean HSI. In the process of compositing the datasets, high HSI values in late December seemed to be disappeared by the low HSI at early December.
VGPM depends on the variables of SST and chlorophyll a concentration, which are overlapped with the parameters used in the HSI model. Furthermore, in this study, SST values were classified into two segments, winter–spring and summer–autumn. The explanation for the influence of this partition on the results of PP values needs to be more explicit.
The variables may appear to overlap each other since the PP was derived with the chl-a and SST data. However, the chl-a and SST data obtained from satellite can measure only in surface, while the modified VGPM - used in this study – is integrated productivity in euphotic depth. Thus, PP, chl-a, and SST are not overlapped variables, and they can all be important variables. Also, the PP was calculated before dividing SST into the two groups. Thus the segmentation would not affect the PP values.
The authors employed 8-day composite images of the Level 3 MODIS data as described in Lines 77–79 of the method section. How did the authors calculate monthly composite images and the HSI values represented in Figure 7? This point should be clearly stated in the manuscript.
The climatological monthly mean HSI were derived by averaging for each month. We revised the sentence (Line 179).
[Minor Comments]
Line 22: SST and SSHA
SST and SSHA should be spelled out when the term appears in the text for the first time.
Lines 111 – 115: Although 37 % of the catch data could not…
This sentence is simply repetition and should be deleted.
Lines 118, 125, 171,173: (Figure 3), (Figure 4) …
Please check the figure number through the overall manuscript and revise it accordingly.
Figure 10
Figure 10 is redundant as this information already appears in Figure 7.
Reviewer 3 Report
General comments
Overall the paper is fairly straightforward, well-written with only minor grammatical errors and I think it is worthy of publication after minor revision The paper needs to be reorganized to include a description of the Japanese common squid, including environmental tolerances, behaviors (feeding/spawning), etc., at the beginning of the article, preferably in a subsection in section 2. Without this information it is difficult to evaluate your choice of environmental variables and the resulting HSI. Most of this information is in the discussion and it needs to be presented at the beginning of the manuscript The paper does not present the time series, or annual/seasonal cycles, of the remotely sensed variables in the study region so it’s difficult for a reader who is not well-versed in this specific area of the ocean to place the optimal ranges into proper context. Climate change is mentioned as a motivating factor but it is not addressed in the results or discussion. Are the effects of climate change evident in the remotely sensed data that you are using to derive the HSI? Is SST increasing? Is the region of suitable habitat expanding or moving away from the native range?Specific comments
Lines 67-68: are the data from the Large Purse Seine Fishery Cooperatives of South Korea available online? How could one gain access to these data? Can you provide any more details about the fishery data? Have they been used in any other relevant publications? I assume these are fisheries-dependent data? Have they been adjusted for effort? See Bishop (2006) Standardizing fishery-dependent catch and effort data in complex fisheries with technology change. Rev Fish Biol Fisheries. 16:21–38. doi:10.1007/s11160-006-0004-9 Line 70 - I assume M/T corresponds to metric ton? Line 77 - can you change the link to direct readers to the MODIS product that you used? The link in your manuscript directs the reader to a listing of all available data from the Ocean Color group. Also, each MODIS dataset has a unique doi, which should be cited Equation 1: Can you describe which dataset or method used to obtain E0, Zeu, and DL? As far as I can see, you only describe PBopt Line 94: can you justify your choice of environmental variables? What about salinity, or mixed layer depth? Why SSHA? Does SSHA correspond to increased mesoscale variability? How does SST compare to subsurface temperatures where I assume the squid are located? Lines 108-113: This material reads as methods and not results. You first get to results on line 116 Line 110: Have considered using an optimal interpolation product for SST to work around clouds? Line 117: does the timing of the catch have anything to do with the lifecycle or migration pattern of the squid, or the fishing schedule of the fleet? Do the fishermen target other species during the other months? Line 123: It’s a bit nit-picky, but squid are cephalopods, not fish. The data are fisheries data, but squid should not be referred to as ‘fish’ Line 124: The ranges for most of these variables are rather large, especially SST and chla, which makes me think that this particular squid is rather robust and does not require specific environmental conditions and can withstand extremes. If this is the case, then the range of suitable habitat is rather large and, therefore, difficult to constrain. How does this impact the HSI? Figure 9: The color contours are the HSI, but what are the black circles? Catch? Please indicate this in the caption Line 186: please describe the optimal ranges for this squid in section 2. We need this information to understand your HSI How sensitive is your HSI to the ranges that you use? The agreement is quite good, but the native range of the species is largely known, the fishermen probably target the migration, biasing the catch totals, and the range of environmental parameters is rather large, as noted above. So, in a sense, you cast a rather wide net. What if you used a different percentile? Lines 188-200: the information about the prey of this squid is very important and this paragraph should be moved to the beginning of the manuscript, in section 2, so those of us who are not familiar with this species can understand it’s optimal environmental conditions and food sources Lines 201-208: same as the previous comment; most of this is background information that is needed upfront and should be presented in the methods or even introduction and not in the discussion Line 205: The Aquamaps site is quite interesting, but it shows the native distribution of T. pacificus to be largely confined to the Yellow Sea, Sea of Japan, and East China Sea. I think you should emphasize that your results reveal interesting seasonality that’s not reflected in established native ranges, like those shown in Aquamaps Also- from the Aquamaps website, perhaps you should include a citation to their publication: Scarponi, P., G. Coro, and P. Pagano. A collection of Aquamaps native layers in NetCDF format. Data in brief 17 (2018): 292-296. Lines 214-216: Can you add the names of the seas to the map in Figure 1? Line 216: Again- migration routes and spawning behavior are important background information that should be presented early in the manuscriptAuthor Response
Overall the paper is fairly straightforward, well-written with only minor grammatical errors and I think it is worthy of publication after minor revision The paper needs to be reorganized to include a description of the Japanese common squid, including environmental tolerances, behaviors (feeding/spawning), etc., at the beginning of the article, preferably in a subsection in section 2. Without this information it is difficult to evaluate your choice of environmental variables and the resulting HSI. Most of this information is in the discussion and it needs to be presented at the beginning of the manuscript The paper does not present the time series, or annual/seasonal cycles, of the remotely sensed variables in the study region so it’s difficult for a reader who is not well-versed in this specific area of the ocean to place the optimal ranges into proper context. Climate change is mentioned as a motivating factor but it is not addressed in the results or discussion. Are the effects of climate change evident in the remotely sensed data that you are using to derive the HSI? Is SST increasing? Is the region of suitable habitat expanding or moving away from the native range?
Yes, the remotely sensed SST shows increasing trends (approximately 0.1 ℃ per year). Moreover, the primary production in the East/Japan Sea shows decreasing trends (Joo et al., 2016, remote sensing). However, the long-term analysis of the suitable habitat area is not conducted in this study. That will be our further study.
Specific comments
Lines 67-68: are the data from the Large Purse Seine Fishery Cooperatives of South Korea available online? How could one gain access to these data? Can you provide any more details about the fishery data? Have they been used in any other relevant publications? I assume these are fisheries-dependent data? Have they been adjusted for effort? See Bishop (2006) Standardizing fishery-dependent catch and effort data in complex fisheries with technology change. Rev Fish Biol Fisheries. 16:21–38. doi:10.1007/s11160-006-0004-9
Unfortunately, the fishery data is unavailable in online. We obtained the datasets from the Korea Institute of Ocean Science Technology (KIOST) during the project. They acquired the datasets from the Large Purse Seine Fishery Cooperatives, and they provided us. The datasets were used in Lee et al. (2018, Remote Sensing). When we tried to standardize the catches, we failed to investigate the favored environmental conditions. With this approach, however, our results for favored environmental conditions of the pacificus showed a good coincidence with optimal conditions for the T. pacificus reported previously. Also, in this study we tried to focus on the seasonal and spatial variations of relative abundance in terms of HSI. Additionally, this study is a kind of pioneer study in the South Korea and we will try to develop a better algorithm for near future. It could be further improved in the next study.
Line 70 - I assume M/T corresponds to metric ton?
Yes, we revised the line 85.
Line 77 - can you change the link to direct readers to the MODIS product that you used? The link in your manuscript directs the reader to a listing of all available data from the Ocean Color group. Also, each MODIS dataset has a unique doi, which should be cited
We revised the link and added references to each dataset (Lines 92, 95).
Equation 1: Can you describe which dataset or method used to obtain E0, Zeu, and DL? As far as I can see, you only describe PBopt
Zeu is euphotic depth, which is derived by the equation [4.6/Kd490] that represent 1% penetration depth of 490 nm radiation (Kirk, 1994). E0 is the amount of incident photosynthetically available radiation (PAR) during the day (E m-2 d-1). Both Kd490 and PAR datasets were obtained from MODIS-Aqua satellite, which are provided by NASA OBPG. DL, or daylength, was computed with latitude and date mathematically. We revised the paragraph in line 100-103.
Line 94: can you justify your choice of environmental variables? What about salinity, or mixed layer depth? Why SSHA? Does SSHA correspond to increased mesoscale variability? How does SST compare to subsurface temperatures where I assume the squid are located?
Salinity was not considered in this study, since its variation in the East/Japan Sea and the South Sea mostly fell within the reported optimal ranges (from 31.93 to 35.7) for the pacificus. We already mentioned about the salinity in line 201-203. In case of mixed layer depth (MLD) dataset, it is hard to use in this study due to its relatively poor spatial resolution (0.5×5 degrees or poorer). We need spatial resolution of at least 0.17×0.17 degrees due to the resolution of the catch data. Thus, the MLD is not considered in this study. The sentence about the MLD was added in line 203-207. Generally, the habitat depth of pacificus is known to be at least 30 m and up to 100 m. It seems to be nonsense that HSI derivation for the T. pacificus with SST data, but the seasonal habitat distribution would be reflected in the range of SST. Also, as far as we know, the subsurface temperature is not a completely separate variable from SST. The sentence was added in line 218-220
Lines 108-113: This material reads as methods and not results. You first get to results on line 116
We moved the paragraph to methods in line 110-114.Line 110: Have considered using an optimal interpolation product for SST to work around clouds?
Yes, we have considered using an OISST datasets. We already know about OISST datasets are really useful and good datasets. However, even if temperature data are available around clouds, the ocean color products (Chl, PP) will be masked. Then the number of matchable catches will be same whether we use OISST or MODIS. Also, there can be some difference when we use OISST and MODIS together in the process of deriving PP (may not be).
Line 117: does the timing of the catch have anything to do with the lifecycle or migration pattern of the squid, or the fishing schedule of the fleet? Do the fishermen target other species during the other months?
As far as we know, most of fishermen rely on the experiential knowledge when planning fishing. Also, they didn’t target other species during the non-fishing season (non- means not zero but low) because the ‘squid catching’ is an occupation for most of fishermen in South Korea. Their fishing is stopped only in a close season for squid.
Line 123: It’s a bit nit-picky, but squid are cephalopods, not fish. The data are fisheries data, but squid should not be referred to as ‘fish’
We revised the sentence in line 139.Line 124: The ranges for most of these variables are rather large, especially SST and chla, which makes me think that this particular squid is rather robust and does not require specific environmental conditions and can withstand extremes. If this is the case, then the range of suitable habitat is rather large and, therefore, difficult to constrain. How does this impact the HSI?
The ranges of environmental variables in Figure 3 seems to be wide. It is also true that pacificus is more robust than we expected. But the optimal environmental ranges where most of the catches (more than 80%) are significantly narrower than total ranges. Moreover, the SST and SSHA ranges are divided into two seasonal groups in Figure 5. Each group shows quite narrower ranges. Therefore, although the squid has a wide habitat range, it is possible to specify the favorite habitat range.
Figure 9: The color contours are the HSI, but what are the black circles? Catch? Please indicate this in the caption
We revised the caption in Figure 9.
Line 186: please describe the optimal ranges for this squid in section 2. We need this information to understand your HSI How sensitive is your HSI to the ranges that you use? The agreement is quite good, but the native range of the species is largely known, the fishermen probably target the migration, biasing the catch totals, and the range of environmental parameters is rather large, as noted above. So, in a sense, you cast a rather wide net. What if you used a different percentile?
The optimal ranges in this study defined as the 10th percentile and the 90th percentile as mentioned in manuscript. We simply compared optimal ranges with known habitat environmental ranges of pacificus. Also, we agree with your comment about the bias of catch data. However, although biases may exist in the catch data, our HSI model can describe seasonal variations of HSI for the T. pacificus. Our next step (or goals) will be the long-term analysis of the HSI with a standardized catch data.
Lines 188-200: the information about the prey of this squid is very important and this paragraph should be moved to the beginning of the manuscript, in section 2, so those of us who are not familiar with this species can understand it’s optimal environmental conditions and food sources
Lines 201-208: same as the previous comment; most of this is background information that is needed upfront and should be presented in the methods or even introduction and not in the discussion
Line 216: Again- migration routes and spawning behavior are important background information that should be presented early in the manuscript
We revised the paragraph in Introduction (Line 41-51).
Line 205: The Aquamaps site is quite interesting, but it shows the native distribution of T. pacificus to be largely confined to the Yellow Sea, Sea of Japan, and East China Sea. I think you should emphasize that your results reveal interesting seasonality that’s not reflected in established native ranges, like those shown in Aquamaps Also- from the Aquamaps website, perhaps you should include a citation to their publication: Scarponi, P., G. Coro, and P. Pagano. A collection of Aquamaps native layers in NetCDF format. Data in brief 17 (2018): 292-296.
The sentence is added in section 4.2. Also, reference was made in the format presented on the AquaMaps terms of use and conditions (https://www.aquamaps.org/main/TermsConditions.php ; General citation; Kashchner et al)
Lines 214-216: Can you add the names of the seas to the map in Figure 1?
We revised the Figure 1.Round 2
Reviewer 1 Report
I have now gone through the comments and revised version provided by the authors, which made a significant improvement-both in scientific and technical aspects as well as in the English language, over the original manuscript. I am therefore much pleased with this new version, which I believe can now be accepted.
Reviewer 2 Report
In the revised manuscript, the authors made a significant improvement. I now recommend publication for the revised manuscript.
