Biodiversity and Habitat Assessment of Coastal Benthic Communities in a Sub-Arctic Industrial Harbor Area

Round 1

Reviewer 1 Report

The text is dense and a stronger trajectory would improve the read.

The biggest concern is the conclusions, which do not adequately summarize the work.  The abstract has related issues.

The paper is very descriptive, it needs a much better defined synthesis at the end. 

Explicate the patterns concerning metals more clearly. 

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The paper of Dreujou et al. assesses macrobenthic community composition in subarctic habitats. They report on the community composition for an understudied habitat and locality, and relate it to habitat variables and heavy metal concentrations. This study represents a baseline for monitoring programs by characterizing the community, as well as offer some initial insights into the effects that an industrial harbor has on community composition.

 

In general, I think that paper is lacking methodological detail, and the logic for the different sampling regimes and comparison of different models needs to be justified. Additionally, the presence of what I assume are program specific acronyms and outputs should be avoided for ease of reading. The model selection needs to be elaborated on, as it’s currently written it is a black box, and the reader is left wondering how the authors arrived at their models.

 

This paper could be improved with a better explanation and justification for the different sampling regimes used and a more clear discussion comparing and contrasting them. As a reader, it was difficult to keep straight all of the different sampling processes and analyses conducted. For example, if I understand it correctly, the 0.5mm size class was sampled only at 2 sectors in 2014 and 2017, whereas the 1 mm size class was sampled at all 4 sectors in 2016. In addition to this, metals were only sampled at 1 sector (Sept Íles) in 2014 and 2016. Finally, in the discussion, all of the models are discussed interchangeably, and it is difficult to remember what data is associated with which model, and if their comparisons are justified.

 

It seems to me that there are two main threads throughout the paper. The first is comparing the community composition estimates based on two size classes, and the second is looking at how habitat and heavy metal concentrations structure communities. To me, the first thread seems to be more of a methodological concern, i.e., when conducting community surveys in this habitat, should you use a 0.5 or 1 mm sieve? The second seems to be assessing how habitat (i.e. benthic substrate) and human activity (i.e. organic content, heavy metals) affects community composition. Both of which are valuable additions to understanding community structure in subarctic habitats, but it is not clear how these different data sets and models can or should be compared. I’m not saying they can’t be compared, but I think the authors need to make a clear case for which comparisons are being made, and the statistical appropriateness of these comparisons. I would recommend that the narrative needs to clearly define the goals of each, and treat them separately (or perhaps they would be better served in more than one paper). Section 3.2 goes in to exhaustive detail about the differences in results when comparing the two size classes, which suggests to me that this may require its own paper, and again brings up concern in the validity of comparing the different models given that they are all constructed on different data sets. Furthermore, the metal content was only assessed at one sector, so making broad statements interpreting the effects of metal concentrations on community composition across the broader sampling region seems unwarranted.

 

Specific comments.

L24-25: the statement “except for some areas in the Baie des Sept Îles where higher organic matter and heavy metal concentrations were detected” is misleading because metals were only sampled in this sector.

L 127: were sediment samples from all sites and years processed?

L134: Spell out UQAR/ISMER acronym. I assume this is a University?

L139: clarify what rare to frequent effects means. I see that this is actually done in the discussion. I recommend just a brief statement clarifying that rare to frequent is a scale relating to mild to extreme toxicity or similar.

L145: It is not clear what is meant by “without mineral or algal material”.

L 156: Define Chao2 estimator. I assume this is a program specific acronym?

L158: I am not familiar with Pielou evenness or taxonomic distinctness (is this similar to alpha diversity?). A formula or brief description would be helpful.

L160: How do these clusters relate to habitat or metal concentrations? – again this is mentioned in the discussion, but including the details of the analysis in the methods is helpful.

L166: were interactions between variables considered?

L167: Was a composite model including habitat and metal variables considered? Or were  habitat-only and metal-only models considered separately?

L 173: The model selection process in the MASS package needs much more detail. Was this forward or backward selection? Or were all possible models fit and the one with lowest AIC score selected? Also see comment below about table 1.

L 176: were only habitat variables used in the distance-based linear modelling, or does this include metal concentrations?

I am not familiar with distance-based linear modelling, or distance based redundancy analyses. Some description of the process here (or in supplementary information if limited by space) would be helpful.

L188: are the Zostera marina meadows a specific type of habitat? Were samples collected directly from here?

L189-191: The values and averages here need to be clarified, and a variation estimate would be helpful. i.e., samples typically had high sand content (52% on average). What is meant by typical? Is the 52% an average of a subset of sites, or all sites? What is the variation SE, SD, etc.? Likewise, some shallow stations had high clay content (83% on average). How many stations had high clay content? Is the 83% value the average of just these sites?

L202: are the highest metal concentrations values here still within the low- to moderate-toxicity levels?

L210-211: It is not clear what is meant by “correspondence levels” here. Does this mean the percent of taxa in the present study that also appear in these reference data bases?

L222-223: I don’t understand this sentence. It is not clear how you can find 137 taxa in common between the size classes, but then only have 114 and 38 in each size class respectively. Unless this is in common to something else?

L237-239: This information may be better presented in a table. An variation estimate would also be helpful i.e. mean (+- SD) or similar.

L244-246: Report the average taxonomic distinctness value and the confidence interval.

Figure 3: I’m not really sure what this figure is showing. Are the curved lines the confidence intervals referred to in line 245?

Table 1: This table could be cleaned up for presentation. Perhaps add row lines to separate response variables. The use of “n.s.” for non-significant is confusing. For example the “Gravel” predictor is included in the models for density and Shannon diversity but is not significant (p > 0.05). I think “n.s.” means it was not included in the model during the model selection process? I would recommend replacing “n.s.” with a “--“ or something similar to indicate it was not in the model. Also see comment above about more details needed for model selection in methods.

It would also be helpful to have a column of n-values and a brief reminder of which data is used in each model:size class combination to aid the reader. For example, I believe the habitat model for 1 mm size class includes all sectors only for 2016, whereas the metal model for the 0.5 mm size class is only for Sept Íles in 2014 and 2017?

L282: what does “all predictors” mean in this context? All predictors assessed, or only all of the predictors that were selected in a specific model?

L286-288: “... which could be the result of sampling strategy…” See previous comments, but this is a major point that needs to be addressed in the methods, or some clear justification for comparing these different data sets needs to be presented.

Section 3.2 seems long-winded. It seems to me that there are markedly different results when comparing the 0.5 and 1 mm size classes. And as the authors note, this may be due to differences in sampling regimes.

L317 - Explain what “IndVal” is.

Table 3 – I’m confused on what this table is showing. Are these values part of a statistical analysis, or were the predictor values for each sample within a cluster averaged together? Does this table include all samples? How were the mean metal concentrations calculated? I.e. metals were not assessed at all sites and in all sample years.

Cluster interpretations – It seems to me that one of the most important insights in this analysis is the collection of taxa as tolerant or sensitive to disturbance in the different clusters, but as written this is buried at the end of the respective paragraphs. If the taxa within each cluster are indeed primarily “sensitive” or “tolerant” to some type of impacts, I would recommend emphasizing this more, particularly when discussing the importance of community surveys in biomonitoring programs. For example, Cluster A has tolerant taxa, which habitat or metal concentrations could potentially explain this? Upon further reading I see that this is done to a degree in the dbRDA discussion.

L350 – the SIMPER routine needs to be explained somewhere

L365 – I believe the statement “… present higher concentration than the rest of our study area… ” is in reference to metal concentrations. If this is the case, it needs to be clear that this is only in reference to sites that were analyzed for metals.

L395 – here it is stated that cluster E has lower metal concentrations, but it is also stated that most of sites in cluster E are outside of Sept Íles, which I thought was the only site where metals were analyzed. Please clarify.

Supplemental

Supplemental figures are hard to interpret when many gray circles are present. Suggest outlining points in black, and/or filling in full circle with color. If you fill them in, decreasing the alpha level may help where points overlap.

Figure s3 c,d – difficult to interpret colors (many are gray)

Table s1 – Adding a column for the number of stations that had a given taxa would be helpful.

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

I have read through the authors' responses and they have adequately responded to all of my original comments. The structure and flow of the paper is much improved, and the methods are adequately described. 

 

Minor formatting note, all of the tables in the pdf I received were not formatted properly (no columns, tables are several pages long), but I assume this was a part of the pdf stitching process. Just mentioning it here to ensure it is fixed before publication. 

Also the second to last paragraph in the conclusion section is repeated.