Dietary Changes in the Ark Clam (Anadara kagoshimensis) Is Associated with Phytoplankton Community Patterns in a Temperate Coastal Embayment

Round 1

Reviewer 1 Report

General comments

The manuscript attempts to investigate the variation of a phytoplankton community and diets of a clam in a coastal system in South Korea. While I think the article is interesting it has suffered from several flaws that may render some of the conclusions wrong. First, I think the mixing models are wrong as the authors used SIAR (instead of MixSIAR) which has better error structures. While I don’t have very huge issues with using an older mixing model approach, my issues are with the use of the Trophic discrimination factors (TDF) which were not justified adequately. Also, the values you reported in your manuscript are standard errors when they should be standard deviation (lines 294-395). As such, your mixing model results are wrong and as such need to be reanalyzed. Second, I don’t know if two sites that are seemingly closed gives enough data to make some of the claims made on phytoplankton dynamics. Perhaps it would be easier if you spoke a little about your samples size at each site. I found myself having to scratch my head to figure out how many samples were collected at each point. Third, the authors give a hypothesis that is not even explored in the discussion. The idea of a hypothesis is to explore it in your results and discussion. Last, I found this paper very convoluted and difficult to follow. May I suggest you move all introductory material (e.g., Lines 96-101) to preceding paragraphs and have objective and your hypothesis as a separate paragraph. Lines 87-108: could much easily adjusted to only include eth objectives and not more literature surveys. Similarly, your results section is too long and hardly matches the aims. May I suggest the authors structure the results section based on the objectives and hypothesis set a priori.

Specific comments

Lines 1-3: The title is too long and winded. Change your title to reflect the findings. For instance, what did your data show about phytoplankton communities and the diets of clams. Also add a species name for the clam.

Line 17: place “Anadara kagoshimensis” in parentheses

Line 30: “deteriorate” I have no idea how the preceding sentences suggest deterioration of food quality. Are you sure you mean deteriorate.

Line 53: add reference for the assertion about human consumption

Line 56: you need to give an idea of what ‘subisidies’ you are referring to.

Line 63: you need to change “the” to “an”

Line 75 – 76: These are unsubstantiated claims – what methodological limitations? What methods are you talking about. You need to add some examples of the methods you are referring to. Also include reference here.

Line 78: topological classes? What does this mean?

Line 119: add parentheses to ‘Phragmites australis’

Line 130: Why the different taxonomic names. You may need to provide references for this.

Line 149: is typical to write it as chlorophyll-a or Chl-a

Line 194: Why is an SOM better than other way of just visualizing community data.  Please add more details about this.

Line 225: Did you test if there was a need for acidification? You need some justification of why you opted to acidify your samples

Line 284-295: SIAR suffers from huge drawbacks especially regarding the error structures. Did you consider using the modern version MixSIAR? Perhaps you can add the iterations and burn-in for your SIAR model.

Line 294: Give justifications for using the chose TEFs

Line 294: The 0.3 is standard error yet SIAR takes standard deviations. You need to convert you values to SD using the samples sizes in the McCutchan et al. 2003 paper.

Line 294: How did you check for model convergence?

Line 301: What are the units for your salinity

Line 316-332: You can pare this down a fair bit.  Perhaps just put the percentage in a table or as supplementary file.

Line 479: Why would your discussion start with your SOM model? better to start by summarizing all results in the order you presented them in the results section.

Line 491-492: What are the implications of these findings?

I did not check references

Author Response

General comments

• The manuscript attempts to investigate the variation of a phytoplankton community and diets of a clam in a coastal system in South Korea. While I think the article is interesting it has suffered from several flaws that may render some of the conclusions wrong.

RESPONSE:

Thanks for constructive comments. Accordingly, we have revised as commented by the reviewer. (see ‘Responses’ to individual comments)

• First, I think the mixing models are wrong as the authors used SIAR (instead of MixSIAR) which has better error structures. While I don’t have very huge issues with using an older mixing model approach, my issues are with the use of the Trophic discrimination factors (TDF) which were not justified adequately. Also, the values you reported in your manuscript are standard errors when they should be standard deviation (lines 294-395). As such, your mixing model results are wrong and as such need to be reanalyzed.

RESPONSE:

Thanks for critical comments. According to the reviewer’s comment, we have re-run the mixing model calculation using the MixSIAR (Stock and Semmens, 2016).

1. To do this, we have used (1) Trophic discrimination factors (TDF) of mean ± SD, which were calculated from values determined for infaunal suspension-feeding bivalves (Yokoyama et al., 2005) instead of common values proposed by McCutchan et al. (2003), and source isotope values recently measured for MPB and Phragmites by one of our co-authors (C. Kim).

References:

- Stock, B.C.; Semmens, B.X. 2016, MixSIAR GUI User Manual, Version 3.1: https://github.com/brianstock/MixSIAR.

- Yokoyama, H.; Tamaki, A.; Harada, K.; Shimoda, K.; Koyama, K.; Ishihi, Y. Variability of diet-tissue isotopic fractionation in estuarine macrobenthos. Ecol. Prog. Ser. 2005, 296, 115–128.

2. We have revised ‘2.7. Isotope Mixing Model’ of the ‘Materials and Methods’ section in accordance with the explanation of the MixSIAR mixing model run. (see Lines 287-302 in the revised ms)
3. We have also revised ‘3.6. Contributions of Organic Matter Sources to Clam Nutrition’ of the ‘Results’ section with modified ‘Figure 8’ in line with the MixSIAR model results. (see Figure 8 and Lines 471-478 in the revised ms)

• Second, I don’t know if two sites that are seemingly closed gives enough data to make some of the claims made on phytoplankton dynamics. Perhaps it would be easier if you spoke a little about your samples size at each site. I found myself having to scratch my head to figure out how many samples were collected at each point.

RESPONSE:

Agree. According to the reviewer’s comment, we have added the following explanation: “On each sampling occasion, duplicate measurements for Chla (and photosynthetic pigments) and SPM at each station were conducted and mean values of individual prrameters were presented in the present study” to ‘2.2. Collection and Treatment of Samples, and Analyses of Environmental Parameters’ of the ‘Materials and Methods’ section. (see Lines 158-160 in the revised ms)

• Third, the authors give a hypothesis that is not even explored in the discussion. The idea of a hypothesis is to explore it in your results and discussion.

RESPONSE:

Thanks for constructive comment. As commented by the reviewer, we have revised the first paragraph of the ‘Discussion’ section to summarize that our initial hypothesis can be supported by our results. Using the 2 subheadings, ‘4.1. Phytoplankton Community Patterns’ was firstly discussed and then ‘4.2. Temporal Variation in the Contribution of Phytoplankton to Clam Nutrition’ was discussed in association with the changes in the community structures of phytoplankton (e.g., dominance of micro- vs. nano- and pico-sized plankton). (see Lines 486-505 in the revised ms)

• Last, I found this paper very convoluted and difficult to follow. May I suggest you move all introductory material (e.g., Lines 96-101) to preceding paragraphs and have objective and your hypothesis as a separate paragraph. Lines 87-108: could much easily adjusted to only include eth objectives and not more literature surveys. Similarly, your results section is too long and hardly matches the aims. May I suggest the authors structure the results section based on the objectives and hypothesis set a priori.

RESPONSE:

Agree and thanks for this constructive comment.

1. As suggested by the reviewer, we have deleted some description on the last paragraph (e.g., Lines 96-101 in the original ms) of the ‘1. Introduction’ section because these descriptions were nearly identical to the descriptions in the second paragraph of the ‘1. Introduction’ section. Then these materials have been retained in the first paragraph of the ‘Discussion’ section. (see Lines 486-505 in the revised ms)

As a result, the last paragraph of the ‘1. Introduction’ section has been condensed just to have objectives and our hypothesis, as suggested by the reviewer. (see Lines 90-106 in the revised ms)

2. We agree with the reviewer. Our result section is a bit long. However, our results are composed of 4 major contents as follows: (1) Environmental conditions that drive phytoplankton community structures, (2) Phytoplankton community, (3) Source discrimination by isotope values, and (4) lam isotope composition and mixing model calculation. It is not easy to condense the results because all these contents are needed to associate clam dietary change with phytoplankton community change, and sustain the logic of the purpose of the present study. So far, trophic research effort presenting invertebrate dietary shift associated with phytoplankton community change has been very rare. Although we have tried to reduce the results section, we have decided to remain as it stands because we need to introduce data for various biotic and abiotic factors.  

Specific comments

• Lines 1-3: The title is too long and winded. Change your title to reflect the findings. For instance, what did your data show about phytoplankton communities and the diets of clams. Also add a species name for the clam.

RESPONSE:

Thanks. Accordingly, we have changed the title as follows:” Dietary changes in the clam Anadara kagoshimensis associated with phytoplankton community patterns in a temperate coastal embayment”. We have also added the species name to the tile as suggested by the reviewer. (see Title in the revised ms)

• Line 17: place “Anadara kagoshimensis” in parentheses

RESPONSE:

We have added parentheses here accordingly. (see Line 17 in the revised ms)

• Line 30: “deteriorate” I have no idea how the preceding sentences suggest deterioration of food quality. Are you sure you mean deteriorate.

RESPONSE:

Agree. We have revised “deteriorate food quality” to “decrease the availability of preferred items (i.e., size-related food quality)” accordingly. (see Lines 30-31; Lines 604-605 in the revised ms)

• Line 53: add reference for the assertion about human consumption.

RESPONSE:

Agree. We have added the relevant references accordingly. (see Line 54 in the revised ms)

• Line 56: you need to give an idea of what ‘subisidies’ you are referring to.

RESPONSE:

As suggested by the reviewer, we have added “(e.g., microphytobenthos and marsh detritus)” here. (see Lines 56-57 in the revised ms)

• Line 63: you need to change “the” to “an”

RESPONSE:

Accordingly, we have changed in the revised ms. (see Line 63 in the revised ms)

• Line 75 – 76: These are unsubstantiated claims – what methodological limitations? What methods are you talking about. You need to add some examples of the methods you are referring to. Also include reference here.

RESPONSE:

Agree. We have added a couple of examples to explain methodological limitations as follows: “because of methodological limitations according to quick digestion after feeding, difficulties in identification of too many taxonomic groups, and often empty stomachs [26]” (see Line 76-78 in the revised ms)

• Line 78: topological classes? What does this mean?

RESPONSE:

According to the reviewer’s suggestion, we have added here “different nodes of food web networks” to help understanding. (see Lines 80-81 in the revised ms)

• Line 119: add parentheses to ‘Phragmites australis’

RESPONSE:

We have added in parebtheses accordingly. (see Line 117 in the revised ms)

• Line 130: Why the different taxonomic names. You may need to provide references for this.

RESPONSE:

We have deleted different taxonomic names because this paper is not relevant to taxonomic identification of ark clams.

• Line 149: is typical to write it as chlorophyll-a or Chl-a

RESPONSE:

Thanks. We have just used “Chla“ as an abbreviation of chlorophyll a throughout the manuscript.

• Line 194: Why is an SOM better than other way of just visualizing community data.  Please add more details about this.

RESPONSE: Thanks. We have added the following sentence accordingly: “The SOM performs a nonlinear projection of multivariate datasets onto a two-dimensional space”. (see Lines 196-197 in the revised ms)

Further explanation for advantage of this methodological approach has been described in the subsequent sentences.

• Line 225: Did you test if there was a need for acidification? You need some justification of why you opted to acidify your samples

RESPONSE:

Thanks. Acidification has been commonly used to eliminate inorganic carbon (calcium carbonates, bicarbonates, etc.) in analyzing organic carbon contents and stable isotope values of organic carbon. In the present study, we also analyzed before (for d¹⁵N) and after (for d¹³C) acidification and found fairly different d¹³C values. Because of a possibility of acidification effect on d¹⁵N values, we analyzed stable isotopes non-acidified samples again.

• Line 284-295: SIAR suffers from huge drawbacks especially regarding the error structures. Did you consider using the modern version MixSIAR? Perhaps you can add the iterations and burn-in for your SIAR model.
• Line 294: Give justifications for using the chose TEFs
• Line 294: The 0.3 is standard error yet SIAR takes standard deviations. You need to convert you values to SD using the samples sizes in the McCutchan et al. 2003 paper.
• Line 294: How did you check for model convergence?

RESPONSE:

These comments are the same as commented above. Accordingly, in addition to applying the MixSIAR and choosing mean and SD values of TEFs (see earlier responses), we have also added the following sentence to the ‘2.7. Isotope Mixing Model’ of the ‘Materials and Methods’ section in relation to the iterations, burn-in, and model convergence: “The model ran Markov chain Monte Carlo sampling based on the following parameters: chain length = 100,000, burn = 50,000, thin = 50, and chains = 3. Diagnostic tests (Gelmin–Rubin, Heidelberger–Welch, and Geweke) and trace plots were applied to examine for any model convergence”. (see Lines 294-302 in the revised ms)

• Line 301: What are the units for your salinity

RESPONSE:

We have added “psu” accordingly. (see Lines 308 in the revised ms)

• Line 316-332: You can pare this down a fair bit.  Perhaps just put the percentage in a table or as supplementary file.

RESPONSE:

Agree. We have reduced some explanation accordingly. (see Lines 329-333 in the revised ms)

• ‘Line 479: Why would your discussion start with your SOM model? better to start by summarizing all results in the order you presented them in the results section.

RESPONSE: Thanks for really constructive comment. According to the reviewer’s suggestion, we have changed whole body of our discussion start by summarizing all results in the order and tried to make sure it through a hypothesis-based description. (see Lines 486-502 in the revised ms)

• Line 491-492: What are the implications of these findings?

RESPONSE: Thanks. According to the reviewer’s suggestion, this paragraph has been all revised in the revised ms and the implications of these findings can be found in the ‘subheading section’: ‘4.2. Temporal Variation in the Contribution of Phytoplankton to Clam Nutrition’.

• I did not check references

RESPONSE:

We have checked again all the reference materials. Thanks.

Reviewer 2 Report

Please see my attached review.

Comments for author File: Comments.pdf

Author Response

Summary of manuscript:

- This manuscript describes the analysis of water chemistry, phytoplankton, and clam shells in attempt to determine clam prey and characterize temporal patterns.

RESPONSE: Thanks for constructive comments.

Specific comments:

• Line 138: “water depth of approximately 3 m at spring tides”. Spring tides versus neap tides are patterns of fluctuation. Depth during a low tide or high tide is more meaningful. Depth during spring tides on average will be the same as depth during neap tides.

RESPONSE:

Agree. We have revised as follows: “Station 1 (S1) has a water depth of 2–5 m and Station 2 (S2) has a water depth of 6–9 m during a low and high tide. (see Lines 135-136 in the revised ms)

• Line 140: Was there no survey done in December 2018? Impressive time series, but I notice the gap for this month in the figures. Was the sampling always done at a certain time of the month or lunar cycle or tidal cycle?

RESPONSE:

Thanks. We have added the following sentence to explain more detailed sampling protocol: “clams were sampled monthly from June 2017 to March 2019, with the exception of December 2019, at the two stations. The samplings were conducted at the end of each month”. (see Lines 138-140 in the revised ms)

• Related to above point, I wonder about the shorter term temporal patterns which may be overlooked or embedded in the data though. Was a finer time scale study accomplished as well? Is that once of the cited studies?

RESPONSE: This is an important comment. Our study aimed to reveal seasonal variability (rather than short-term variability) in phytoplankton community and clam diets in relation to recent warming of the sea and decline of clam production. Therefore, we have analyzed monthly patterns.

• Line 146: 180um refilter. I am wondering why this was done, some particles larger than this may be contributing to suspended particulate matter. I can understand removing any large zooplankton, but keeping rest of water sample would seem better. Abandoned appendicularian houses, marine snow, might be larger than 180um.

RESPONSE:

Thanks. This is a critical comment. We tried to analyzed stable isotope values of primary sources of organic matter in line with the aim of the present study. As commented by the reviewer, abandoned appendicularian houses and marine snow larger than 180 mm could include debris of consumers, not producers. Therefore, to acquire stable isotope values of producers (i.e., phytoplankton and benthic diatoms, etc.), we filtered water immediately after sampling through 180 mm mesh and tried to extract them.    

• Line 226: For phytoplankton sample this is where I might have filtered out the larger nonphytoplankton but no mention is made. A towed 45cm net would certainly capture some nonphytoplankton, how were these handled?

RESPONSE:

Agree. We have added the following sentence to clarify filtering out procedure of the larger nonphytoplankton: “After filtration using a 180 mm mesh screen,”. (see Lines 231 in the revised ms)

• Line 258: I am wondering if other covariates were examined such as tidal cycle, hour of the day, sunlight/clouds, rainfall, currents, etc… other things more reflective of the local and immediate conditions for a highly dynamic and productive area where the ecosystem could likely respond very quickly?

RESPONSE: Agree. In the present study, we did not have available data for these environmental variables. Instead, we have discussed the importance of tidal exchange between inner and outer-bay water as one of the most important factors in determining the phytoplankton community composition.

• Line 262: A mix of parametric and non-parametric tests are mentioned in one sentence. Was their use dictated by the test of normality? This is implied but not stated.

RESPONSE: Thanks. We firstly tested the normality and homoscedasticity of the variance for all kinds of data. Then, when we met the assumption of normal distribution and equal homoscedasticity, we did parametric tests. When we did not meet, we did non-parametric tests. These procedures were described in the paragraph of the ‘Materials and Methods’ section.  

- Line 469: Clam not calm.

RESPONSE: Thanks. We have corrected accordingly. (see Lines 475 in the revised ms)

 General comments:

• I am impressed with the data collection, and the very nice time series of data. I appreciate the data analysis and care to interpretation; however, I am struck by the apparent lack of match of the clam data to any of the prey options. For example, nothing comes to the level of the 15N values of figure 7. This makes me think that something is missing, that the actual prey was not one of the things analyzed. What could this be? Microzooplankton? Bacteria? I would have liked to see this discussed further in the paper. Certainly someone has studied the clam diet directly, what shows up on their filtering structures or guts? Recommendation: This study presents much useful information despite not leaving me with a satisfying conclusion. I believe it is publishable with a little more text to get at what the clams are eating or not.

RESPONSE:

Completely agree. We aimed to highlight diet change of the clams in association with seasonal shifts in phytoplankton community structures. d¹⁵N values were used in the isotope mixing models. SPOM includes pico-, noan- and microphytoplankton, bacteria, and even microzooplankton, etc. SOM will be a mixture of various origins of organic matter. Therefore, in the revised version, we have reduced source materials into SPOM, microphytobenthos, and Phragmites to do isotope mixing model because microphytoplankton can be embedded to SPOM. This gave rise to more robust resolution. Microzooplankton and bacteria would have a trophic role to the clam diets. However, this is likely to be beyond the scope of the present study because stable isotopes of microzooplankton and bacteria could not be analyzed in the present study.

Round 2

Reviewer 1 Report

The is my second review of the above manuscript. I applaud the authors for the reanalysis using MixSIAR and the clarification of many of the parts of the manuscripts. I still have my doubts about the length of the results but I am not too perturbed about it.

Title: Better. Just change to "Dietary changes in the ark clam (Anadara kagoshimensis) is associated with phytoplankton community patterns in a temperate coastal  embayment

Terms:  When you say Chla - may be better to write is as Chl-a

Author Response

• The is my second review of the above manuscript. I applaud the authors for the reanalysis using MixSIAR and the clarification of many of the parts of the manuscripts. I still have my doubts about the length of the results but I am not too perturbed about it.

Response: As indicated by the reviewer #1, We have tried to condense the ‘Results’ section. Finally, we have decided to keep it that way to describe detailed results.

• Title:  Just change to "Dietary changes in the ark clam (Anadara kagoshimensis) is associated with phytoplankton community patterns in a temperate coastal embayment.

Response: As suggested by the reviewer, we have changed the title accordingly. (see Title in the revised ms)

• Terms:  When you say Chla- may be better to write is as Chl-a

Response: Thanks for acceptance.

 Additionally, we have changes Figure 1 and Figure 4 to match the station names in the figures and descriptions. (See Figures 1 and 4 in the revised ms)