Acclimatization Drives Differences in Reef-Building Coral Calcification Rates

Round 1

Reviewer 1 Report

This manuscript shows through reciprocal transplants of two species between two sites that growth and accretion rate responses are distinct between the two species, with Montipora capitata being more stable and accretion rate changes detected for Porites compressa, in particular transplants not accreting as fast as residents at one the sites. The experiment is conceived thoughtfully and looks to be executed well. The analyses are fine though there is concern about the inference of based on the analyses. Below are some comments for consideration.

 

Growth and accretion are overlapping as acknowledged by the authors. I suggest measuring (converting to) change in skeletal density to remove the growth factor from the accretion rate. This will entail measuring the total enclosed volume. Otherwise, I cannot see how accretion rates can be independently interpreted.

 

I am also concerned that the point about plasticity is more prominent that is warranted (i.e. in the title) given that the authors did not test the genotype x environment interaction directly. Genotype (or parent colony) should really be a fixed effect in order to test for the interaction. The large variation attributed to parent colony as a random effect does not explain much since the variation may not be interacting with the environmental differences.

 

Furthermore, as no genotyping (e.g. microsatellites, SNPs) was done, there is no way to tell if there are genetic differences among colonies, how distinct the colonies are genetically and whether these differences are associated with the responses detected, so this needs to be made clear.

 

Tables 3 and 4: Which effects are significant? What are the p-values?

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

page 2 line 57: please, can you be more specific and precise which climate warming scenario you are referring to ? IPCC scenarios ? Which RCP scenario in that case? Do you mean global predictions or local predictions of climate change for Hawai only ? Do you have any clear evidence (please provide a reference)? The global warming issue is such commonplace in ecological studies nowadays, that precise, referenced data are very welcome.

 

Page 3, line 119 : here, it is not totally clear to me how 6 fragments of 5 colonies make 60 samples for each species in total ? Did you miss to mention any replicate, or I may have been mistaken ?

 

Page 5, lines 176-178 : it seems that some differences in nutrient values between the two sites are important (ex: nitrogen, silicate, and ammonia), as suggested in the discussion section in lines 297-298. Could you test for these differences (not tested according to table 2)? If not, please explain why.

 

Page 6, line 204 and following, plus table 2 : it is not clear here, whether differences in accretion rates between sites are significant or not for each species. Can you please precise and indicate in the table (with an asterisk for instance) which differences are significant or not,  if any ? Can you also confirm that differences between transplants and residents are not significant ? This is important for the rest of the analysis.

 

Page 7, lines 212-214 : please explain why you choose to consider accretion rates only. Even if it is correlated with linear extension rates, you may also have chosen to analyze extension rates only. Why did you finally choose to focus on secondary calcification only ? Is it more representative of coral health than primary calcification ? This is important as, as mentioned twice in the ms, correlation (R^2) between the two calcifications is far from R^2=1 and different metabolic responses to environmental variation may be expected for each calcification process.

 

Page 7, line 223: please can you precise what "marginally higher" means. Is the difference significant or not ?

 

Page 7, line 226: similarly, did site 2 transplants accrete significantly more than site 1 transplants ? Figure 4 suggests that there is not difference between transplants of both sites.

 

Page 9, line 262: as you mention that site 1 experienced 2 days with warm temperature peaks above the upper thermal threshold for corals, can you confirm that in contrast, temperatures at site 2 never reached the threshold value ? (if I get it correctly) ? This seems to be more important than the absolute temperature value itself.

 

Page 7, line 265 and following: do you suggest that salinity is an important factor of environmental variation between the two sites? In particular, salinity values at site 2 can be lower than the threshold levels for coral reefs (23.7< 25 ppt) and not always within the coral average range (25-45ppt). This seems to be of major importance for coral survival. Why don't you give much importance to salinity variation in your interpretation, compared to temperature ?

Author Response

Please see the attachment. 

Author Response File: Author Response.docx

Reviewer 3 Report

This is a very interesting and important paper, and I enjoyed reading is for most part.  Most of my comments should only require minor revisions, but I do have a concern regarding how the formal statistical analysis was done.

I understand the difficulty of getting a permit that involves killing corals, but looking at Figure 4 makes me wonder in there were not enough replicates to detect the Site x Treatment interaction for M. captitata (perhaps power analysis can tell...). It is also difficult to tell if the lmer analysis was done appropriately as the manuscript does not describe it in details. It just leaves me wondering if the dataset, which I assume required lots of work to collect, was fully and appropriately utilized.  Please see below for specific comments.

Introduction

Lines 67-69. "The northernmost 30 meters of the reef had 38.9% coral cover, and 58.3% algae cover while the southernmost 30 meters of the reef had 62.5% coral cover and 25% algae cover [Barnhill, unpublished data]." I cannot understand what you mean by the northernmost 30 meters and the southernmost 30 meters.  Is there any way to add this to Figure 1 or please clarify.

 

Methods

Lines 91-92. "Reefs in the Bay have one of the highest percentages of coral cover (54-68%) among the Hawaiian Islands (Hawaii average = 24.1%) [14,15,25]."  Those numbers seem high. Could you clarify how the numbers were calculated (e.g. depth range and only hard bottom habitats or all available areas at the depth range were used for calculation etc.).  Does the number for "Hawaii" include the entire Hawaiian archipelago or just the main Hawaiian Islands (i.e. excluding the NWHI)?

Lines 128-129. "One nubbin per colony was placed on each tray to avoid pseudoreplication." So, each tray had 20 nubbins, one each from 5 different colonies of 2 specie per site, correct?  How did you design the placement of 10 nubbins, any randomization? Please clarify.  Also, how far apart were the 3 trays within each site?

Line 152. It is R that needs to be given credit here, not Rstudio (though you can off course cite Rstudio as well), and it seem the version number (3.6.3) listed is really for R.  Rstudio is just an environment for R.  You can run the function citation() to see how to cite R properly and give well-deserved credit to the R core team.

Lines 158-163.  Could you give more details for this lmer analysis. It was run for each species separately, and the 2 fixed terms were crossed, correct?  How did you include Colony as a random factor?  Was it nested within Site AND Treatment (e.g. Site 1 x Resident combination gets colonies 1-5 and Site 2 x Resident combination gets colonies 6-10 etc)? There is also the Tray effects to consider, which you should include as a random factor nested within Site, as I assume that the colonies on the same tray were much closer together than the colonies on the other 2 trays. Please include it or offer justifications for not including it in the analysis.

Results

Lines 182-184. Could you at least include these outliers in the descriptive statistics and visualization?  Was it just the accretion rates that were unusual, and not the linear extension rates?  You can always include descriptive statistics with and without outliers.  Also, removing outliers just because, well, they are outliers, is not really a good practice, especially when you only have 5 data points per treatment combinations (if you consider the tray effect)... Did you consider non-parametric approaches (e.g. permutational analysis) or data transformation or even using non-normal distributions for your analysis?  Could you offer some justifications for removing them?

Figure 3. Could you also add Site to this figure, perhaps using different symbols?

Lines 212-213. "Due to the strong correlation between linear extension rates and accretion rates (Figure 3), linear extension rates were not included in the analysis." Looking at the spread of the data, I strongly suggest you run the analysis on both. I am not a coral expert, but based on the way they are measures, one looks at more of 2D growth and the other 3D (volume), so they are not the same type of metrics, correct? If results tell the same story, you can always move one to the supplemental materials.  Also, if the outliers were in accretion rates and maybe not in extension rates (not sure since the manuscript does not say), then it makes much more sense to use the full dataset for linear extension rates for this analysis assuming that you had a very strong reason to pick one.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

Thank you so much for all the clarifications.

Unfortunately, I am still a bit confused about how the analysis was done, and my confusion likely comes the the word "Site."

In this study, Site can mean either the origin or the destination of coral nubbins, and they do not seems to be well distinguished in the manuscript.

For example, in the statistical analysis description, it says "The linear mixed effects model was run for each species and each calcification measurement seperately. The mixed effect models used site, treatment the cross between site and treatment as fixed factors. Parent colony was included as a random factor nested within site (Site:Parent_Colony) to control for multiple nubbins being cut from the same parent colony. Tray nested within site (Site:Tray) was set as a random effect to control for the effect of multiple nubbins being placed on the same tray."

Here, I first assumed Site is the destination of nubbins, and Treatment is resident vs. transplant.  So if the Site is the destination, it makes sense that the Tray was nested within Site (Site 1 gets trays 1, 2, 3 and Site 2 gets trays 4, 5, 6, etc). However, it does not make sense that the Parent colony was nested within Site, as Parent Colony should be nested within the origin, not the destination (i.e. colonies 1-5 came from Site 1 and colonies 6-10 came from Site 2). Since Parent Colony was included in the formal analysis, I assume (and hope) Site meant the origin of the colonies, not the destination...

So, please clarify the origin vs. destination throughout the manuscript.

I also suggest using the destination as a factor instead of Treatment (i.e. resident vs. transplant). This shouldn't change the final conclusion (again, assuming Site meant the origin), but it seems to make more sense to test the origin (genetics) and the destination (environment), rather than whether colonies stayed vs. moved from the origin. They are both fixed factor with 2 levels (Site 1 and Site 2) and you can include their interaction (like you did with Site and Treatment). Then you also should be able to test for the random effects appropriately (Parent Colony nested within Origin, Tray nested within Destination). It is a suggestion, but I believe it makes the interpretation of statistical analysis a lot more straight forward.

See below for a couple of super-minor comments.

Lines 69-71. "Benthic cover within 30 meters of Site 1 had 38.9% coral cover, and 58.3% algae cover while the cover within 30 meters of Site 2 had 62.5% coral cover and 25% algae cover [Barnhill, unpublished data]." Please add reference to Fig 1. Otherwise, Site 1 and Site 2 come out of the blue...

Line 81. "as [21] suggests basing coral reef health on reef-building capacities of the corals" I suggest adding the authors' names before [21] so it read "as XXX et al. [21] suggests basing..."

Author Response

Please see the attachment

Author Response File: Author Response.docx