All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
Thank you for addressing all reviewer comments. I have reviewed the manuscript and believe it is ready for publication.
[# PeerJ Staff Note - this decision was reviewed and approved by Anastazia Banaszak, a PeerJ Section Editor covering this Section #]
no comment
no comment
no comment
The authors have sufficiently addressed my concerns and editorial suggestions.
In general, the reviewers appreciated the work and agreed it had merit. They provided several good suggestions, especially regarding the writing (with a focus on the methods) and improving figures. A lot of the comments are around being more precise in the language for describing methods and results.
**PeerJ Staff Note:** Please ensure that all review, editorial, and staff comments are addressed in a response letter and that any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate.
**Language Note:** The review process has identified that the English language must be improved. PeerJ can provide language editing services - please contact us at [email protected] for pricing (be sure to provide your manuscript number and title). Alternatively, you should make your own arrangements to improve the language quality and provide details in your response letter. – PeerJ Staff
I think the work is very good, with clear and novel information. The temperature analysis seems very good and key to understanding the impact of El Niño 2023. The methodological approach is adequate, and the analyses and interpretations are very good. Throughout the work, there is a very strong emphasis on active restoration of the area's coral reefs, without really considering other alternatives. There is a final paragraph on this point, but it is biased to indicate that there are no other alternatives. Restoration is undoubtedly viable, but it can be accompanied by other actions, such as examining the impact of herbivores on the increase in algae. As an MPA, the system itself can aid recovery. While it is true that temperatures have been rising over the last 40 years, it is undeniable that we must seek innovative conservation alternatives, but it is also key to present the full spectrum of these alternatives.
novel and well developed
concrete and good results, with a high management value for the area
See several comments throughout the document. These are several questions, nothing serious, that I hope may be valuable to the authors.
Pass
Requires major revision
Requires major revision
Review of Palmer et al, PeerJ, El NiÒo-Driven Phase Shift to Algal Dominance on Isla
del Cano’s Coral Reefs: Implications for Urgent Restoration
Palmer et al present an interesting dataset examining coral communities in Costa Rica which were affected by the 4th Global Bleaching Event. The dataset incorporates some long-term surveys and hence a variety of survey types, which are incorporated all together. Importantly, these reefs were in the early recovery phase when a landslide delivered mass sedimentation.
While the data seem robust, I have concerns around the analyses. The manuscript is very long, partially because it incorporates a LOT of analyses, many of which could be simplified and collapsed into a single analysis (e.g. benthic cover could be analysed in a single model, rather than coral cover on its own, algal cover on its own, and could include the fixed factors of site and MaxDHW). Additionally, some of the fixed factors should be removed due to collinearity and confounding – i.e. site is very directly linked to temperature and both of these are linked to depth. Simplifying analyses would provide a more compelling, direct, and clear indicator of the changes observed at these sites.
I am particularly unclear on the use of SIMPER, which the authors state identifies the drivers of change in community composition. SIMPER is used to determine the taxa with the greatest relative contribution to change between timepoints, but has no ability to provide ecological inference about the drivers.
The ecological recovery feasibility index is an interesting concept, but in the current manuscript it is under-developed. It’s not clear how the index is calculated or why. Surely restoration success will be dependent on the type of restoration attempted – but this is not discussed – and how to quantify that in a single index is unclear.
More specific comments below:
The abstract is very long- please consider shortening to retain only the most critical information. In particular, the methods are very long.
Line 87 – please capitalise IPCC
Line 90 – ecological collapse can occur even without functional extinction, consider rewording.
Line 93 – do you have a reference for ecological collapses occurring during El Nino events? I would suggest that collapse is a more chronic process, while El Nino events are acute contributing factors.
Line 115 – you already defined/used the abbreviation TEP above, can eliminate here.
Line 115 – I’m curious why these reefs have been purported to be resilient to climate change? What about them underpins this proposed resilience? Coming back to this comment after reading the next paragraph, I see you suggest that survival post 1998 and 2015 events suggests greater tolerance etc. Suggest including a few words in the last line of previous paragraph indicating this; however consider also that it may not be resilience, but instead a selective sweep leaving only the hardy/robust species
Line 130 – there seems to be a grammatical error with wording surrounding the reference
Line 159 – you already defined IPCC, use only the abbreviation here
Methods
Line 192 – You already defined the bleaching threshold above. Remove here. I also suggest moving the “cooling threshold” definition above to where you state the bleaching threshold information.
Line 206 – how were “recently” dead corals defined and identified? Why is Caulerpa its own category?
Line 208 – your supplementary table says n=3 surveys in 2021 – please clarify
Line 210 – how many 1x1m quadrats per 30m transect?
Line 213 – I am unsure of your methods with the LIT. If you recorded every 10cm, isn’t this a PIT? Please clarify. Unsure what you mean by recorded every 10cm and then the continuous 10m line?
Line 214- typo Tn
Line 278 – I question your treatment of missing values. If values were missing for a given benthic category, did this mean it was not recorded in your surveys? If that is the case, why then artificially assign the mean? I may not be understanding correctly what you did – please clarify.
SIMPER analysis – I skipped ahead to read your results at this stage to see if I could follow what you meant about the missing values. However, a different concern was raised for me. I am not sure SIMPER analysis is the most appropriate choice to determine the contributors of benthic change – it isn’t quite accurate to say that turf algae and dead coral were the “drivers” of benthic change – they are the change itself. I believe using a GLMM would be more appropriate (benthic category ~ time * site; or similar), and you could either group your data or make planned contrasts using emmeans to compare the “during bleaching” and “after disturbance” timepoints. However, if you are interested in drivers, then environmental variables should be examined.
Line 288 – unclear wording “with variations in the treatment of time as a predictor.” Does this mean you tried the detailed models and then chose one using AIC or other model selection steps? Coming back to this comment after reading the next paragraph – I think you can omit details on the models you did NOT use and just explain the models you did use and that these were selected via robust model selection and validation steps. Your manuscript is quite long, so these extra details aren’t required.
Line 305-310 – unnecessary detail. Recommend reducing this entire section.
Line 312-316 – I imagine that depth is confounded with site. Is this analysis necessary?
Line 330-346 – What does the HOBO data add that your CRW/DHW data haven’t captured? My suggestion is that the HOBO data could be used to validate the satellite-derived data, but then choose only one for your analysis. MaxDHW might be worth investigating.
Line 366 – I’m unclear what you mean by these factors being absent in the latest analysis. What is the latest analysis? The one included in the manuscript? What is the earlier analysis where the factors Caulerpa and dead coral were not excluded? Unclear.
Line 373 – this is the first time we hear about the landslide. Need to provide relevant background
Line 378 – you can not make direct comparison between different PCAs as they are calculated on distance matrices that relate directly to the samples involved in the analysis. If you wish to make direct comparisons between the during bleaching, after bleaching, and after landslide timepoints, you must analyse these within a single PCA.
Also, suggest using more informative group name than “Latest Data” such as “recovery” or “post-landslide”
Line 388-402 – I’m unclear what the ecological impact directions are. Is this a numeric variable? I see that they take into account different numeric factors, but how these are combined in a calculation is missing. This is a novel indicator, so warrants more significant explanation.
Results
Line 415 – suggest including the actual DHW value for each event
Line 454 – suggest including actual prevalence values for each genus in brackets
Line 461 – grammatical error? and Other coral had high dispersion of Other corals,
Line 462 – zero-inflation does not confirm a match between your model and your data. You need R-squared values for that.
Line484-492 – as above, I am concerned by the wording here, as the taxa themselves can not be the driver of change. They can be quantified as taxa that contribute most to between-timepoint dissimilarity using SIMPER. The wording is important and needs to be clarified. As above, I think a GLMM is likely a better fit for analytical method. The paragraph mostly describes % cover (I believe – or are these the % of observed dissimilarity between timepoints??), so the paragraph subheading is not appropriate. Standard errors are missing throughout.
Lines 484-507 – unclear what test the p-values come from, and their associated test statistics. Are the % cover estimates model estimated means? Where are the standard errors?
Line 502 – “dead coral nearly disappeared” is a strange phrase. The dead corals didn’t disappear, they likely were overgrown by turf algae or other benthos.
Lines 484-507 – these are some of your most interesting results, but we don’t get to see them graphically displayed. I highly recommend including a line graph for the percent cover through time for each benthic category
Line 511 – why use SD and not SE?
Line 512 – SIMPER can not identify a decline in coral cover over time – it can only determine which groups contribute to between time-point dissimilarity.
Line 525 – not a full sentence
Line 545 – depth, site, and temperature are highly confounded
Line 574-636 – the PCA should be completed as one analysis, then separated for visualisation only. As is, it is not valid to make comparisons between different ordinations for each timepoint. If the analysis is redone and still supports greater algal dominance in post-bleaching timepoints, this should still be supported by above recommended analyses of benthic cover through time.
Line 584 – “disappearance of dead coral” again is strange wording
Line 585 – “Latest Data” should be given a more informative group name
Line 594 – if you want your PCs to distinguish particular groups, you could do a constrained RDA
Line 629-636 – it is not accurate to say that loading values reveal declines in coral cover or changes in benthic community – you need to use your absolute abundance/cover data for that, and/or a PERMANOVA to provide statistical support for this statement. You can only make observational statements derived from your PCA about the association of taxa with different groupings.
Line 649 – notwithstanding significant clarification regarding this index (see above comments), there is odd phrasing: “These results highlight that sites with higher feasibility
scores tend to maintain greater coral cover” → didn’t you make the index such that higher feasibility scores were given to sites with greater coral cover? This is a circular argument. I am not sure that high coral cover necessarily equates to greater restoration potential – greater coral cover could also mean increased coral-to-coral competition. Need far more in methods to describe this index and how it is created/used.
Figures:
Figure 2 – move to supplementary. Change white line – hard to see, and is absent from legend
Figure 3 – move to supplementary.
Combine Figure 4, 5, and 6 into a multi-panel figure
Figure 7 – It would be helpful to show your bleaching data alongside temp/DHW data to give context. I highly recommend standardising the X-axis so the reader can see which sites have longer survey durations than others. Can reduce the point size.
Figure 8 – recommend matching your colours between figures
Figure 9 – this is the main event of your paper – though I think figures for other benthic components would be helpful. For this figure, I highly recommend standardising the X-axis so the reader can see which sites have longer survey durations than others. Recommend including an R2 for each panel.
Figure 10 – labels are very hard to read, and often overlap site names, some of which go off the edge of the plot space. If combined into one analysis as suggested, you could use 95% confidence ellipses to show the change in community composition among the three time points. It might be the submission platform that reduced quality, but the resolution is showing very poor.
Discussion
Line 775 – the landslide is really important context for the study, but hasn’t been given sufficient introduction. It is significant that these sites (some of them, at least) were subjected to record-breaking heat stress followed by an acute sediment smothering event. I think the authors could frame their work differently to really focus on these consecutive impacts.
Line 781 – not a full sentence
Discussion is very long, consider your main points and focus in on them. The taxa-specific results could be significantly shortened and simplified. There is a lot of repetition of main take-home points; I recommend spending some time on re-structuring the discussion with clear topic paragraph structure.
Line 875-888 – given the increase in algal cover and the focus on low-tech/low-cost restoration, I think some consideration of manual algal removal would be warranted. See work by Tanner 1995, McClanahan 1999-2011, Bulleri 2018, Smith 2022/2023.
Line 890-898 seems like a departure. How would immune profiling be low-cost? If you profile the immune function and it is compromised, how do you improve it?
The article is largely clear with good use of English, has sufficient references, follows a professional structure, and is largely self-contained with relevant results to hypotheses. See my specific comments for edits.
The experimental design is sound. My overall criticism is that the article appears to do a shotgun of analyses and it isn't always clear why specific analyses are being conduct and to answer what specific question. Some of the analyses seem a bit disjointed from each other, or at least the connections are not that clear. I think the paper could be honed down a lot to very specific objectives. I get the sense that they got a bit "R-happy" without a firm sense of what they wanted to test a priori. Thinking about what you want to test with the data beforehand and crafting the analyses and the framing of the analyses to that would make the manuscript read much better. I think it would also invite a much wider audience, as I think some people might lose patience trying to figure out what is important. In essence you have a thermal stress and bleaching story at multiple sites and can test community response and suggest management implications.
I think the overall conclusions could be a little more tied to the overall objectives of the study. For example, there is a lot of interjection of management implications into the biological outcomes of the study. I get the sense this a clear mission of the authors, but it does not necessarily flow from the biological results per se. I would keep the biology and management separate in the discussion to keep the focus on each topic in turn.
Summary
Background.
Lines 45-47. Confusing sentence. Makes it seem like evaluating the impact of the El Niño is supporting ecosystem services. Also, I think you could present the 2023-24 El Niño as an event first, then talk about ecosystem services. Seems like it needs an introduction.
Methods.
Lines 52-54. DHW and coral cover statistics are brought up before the benthic cover methods are discussed. Perhaps move to the end in a succinct section on all the statistics.
Results.
Lines 68-70. The percentages reported appear to be relative percentages which are somewhat uninformative on their own. E.g., a 40.44% decline in coral cover could represent a change from 2% to 1.2% or a change from 50% to 30%, which have very different biological implications. Maybe you can report as the change in absolute cover. E.g., “Coral cover declined from 50% to 30%.”
Introduction
Line 85. “target temperatures” seems loosely defined. You mean thermal and/or bleaching stress thresholds?
Line 87-90. Change in list structure. Change to “compromising” and “threatening”
Line 111. “garnered the”
Line 123-125. You state no coral recovery between 1982-83 and 1997-98, with values of around 10%, but later in the paragraph you state that current coral cover is 20-70%. Can you say that there was some recover since 1997-98?
Line 162. Delete the “a” before “a fine-scale”
Line 164-5. Delete “With such information,”
Line 170. Keep it past tense. “The aims of the study were…”
Methods
Line 192-3. Can you define the “cooling threshold” more and its purpose in the study?
Line 205: maybe avoid using the word “variable” to not confuse it with statistics. Use “different”
Line 208: Explain how transects were laid haphazardly more. How much error do you think was incorporated into the time series from examining different areas of the reef?
Line 223-4. Explain the statement “though remained temporarily and variably functional” a little more. I don’t understand.
Line 274. What do you mean by “dead coral”? This variable is declining in the post bleaching surveys, but intuitively wouldn’t it make more sense that dead coral increases after bleaching related mortality? Therefore, be very specific about what you mean by dead coral and how this was evaluated. You may also want to make some sort of clarification in the results in the “Drivers of community composition change” section. If you mean recent partial mortality or recent mortality that seems to be a more standard way to define bleaching caused mortality.
Line 312-6. Is this assessed across the entire data set with all sites? Is the depth of each transect taken independently or is the location mean depth used for all transects at that site? This could have big implications for the interpretation of the results and should be carefully reported. Alternatively, I don’t know what this analysis adds in particular to the study, so it could also be removed.
Line 329. Coral cover and local temperature. The methods here are very unclear and it makes it very difficult to interpret the results. Are you comparing the mean coral cover at a site (across time?) to the mean temperature. A lead in sentence that states the purpose of the analysis would be helpful, here and throughout the methods.
Line 387. I am not sure I could replicate the Ecological Recovery Feasibility Index based on the methods provided. Can you provide at least an example of a calculation? Could be in the supplement.
Results
Figure 2. It would be helpful to reiterate how you calculated the DHW from MMM + 0.5°C, since in the caption since this is a less common metric than MMM + 1.0°C. Also provide the equation for the trendline in the caption for easy reference and point out the rate of increased per year or decade.
Coral Diversity
Line 468. Misspelled “Pavona fronidera”
Local impacts, bleaching prevalence and coral susceptibility
I think it would be valuable to evaluate your coral bleaching estimates with respect to the estimated DHW at the time of sampling. This would be more informative than time. You could use either the CRW data or the local data where you have at specific sites. At the very least indicate how close your survey date fell to the peak of heat stress in 2023-24 and how this might have influenced the prevalence of bleaching seen and the resulting coral cover changes and community responses.
Figure 7. Indicate what each data point at each site represents. I presume these are individual transects but this should be explicit.
Drivers of community composition change
Somewhere you should report the absolute values of cover for each of the categories to put the changes in context. This can be a table in supplemental, but I think it might warrant a display in the main paper as either a figure or table.
Depth and site temperature and coral cover
See my comments in the methods. You statement “accelerated decline in coral cover at deeper sites” makes it sound like the comparison is on the pattern of cover over time, but I think you measured just the static mean cover on a transect? The methods and analysis of depth and cover are very unclear and hard to interpret as presented.
Do you think Tina had a higher temperature because it was shallower? Is the peak of coral cover at 8m and the higher temperature correlating to higher coral cover just saying the same thing? Sites with cooler temperatures were deeper and also had lower coral cover.
Also, we need to see this data, not just the statistical outputs. Did you try to calculate DHW for the in situ data using your MMM +0.5°C threshold? That might also be interesting especially in regards to the discussion of lower initial heat stress and coral bleaching at the deeper sites (lines 741-2).
Ecological Recovery Feasibility Index
How were PCA loadings incorporated into the index? I am not entirely clear on this and might use some better description in the methods.
Figure S4. Define the x-axis. What is the unit of time? How is 0 defined. You can include this information in the caption.
Line 543. “Temporal trends were not significant ß = -0.035, SE = 0.033, t = -1.08, p = 0.282; Fig. S3).” I think you mean Figure S4?
Discussion
Line 658. Maybe avoid, or at least define, “reef collapse”.
Line 674. What is a “coral reef critical environmental threshold breach”
Line 684-7. Would read better as “With the exception of the 1983 El Niño, Isla del Caño’s …”. And separate out the last part as it doesn’t read correctly. Maybe a new sentence “However, examination of four decades of SSTs show that thermal events locally were not prolonged or extreme”
Line 698-70. “being outpaced at Isla del Caño,” This is not entirely clear. You mean that you do not expect recovery within the return time of thermal anomalies? Rephrase.
Line 725-7. It is okay to talk about Alternative Stable States and positive feedbacks in relation to you sites, but since you cannot prove this without a lot more data than you have from your study do not make claims that your system exhibits an alternative stable state. E.g., “The progressive shift in reef composition, and the absence of recovery, signifies coral reef degradation and the establishment of an alternative stable state at many sites, with reduced structural complexity and algal dominance (Fung et al., 2011).” Everything in the statement is true, except that this signifies an Alternative Stable State. Also, I do not think post-disturbance recovery period is nearly long enough to really say much of anything about all but the most immediate recovery. Most of this section is about external definitions of reef systems in states of degradation and only tangentially about positive feedbacks. So why not just get rid of the alternative stable states discussion altogether?
However, I do find it interesting that turf cover increases more than it appears that stony coral cover declines. That could be soft evidence that something else has changed besides a 1:1 changeover of benthic space. If you look at the data, does it appear that way to you?
Line 741: “structurally intact”. What does this mean? Be more specific. What in the data indicates they were different from shallow reefs? Level of recent partial mortality.
Line 750-751. “both exhibited increasing alignment with turf algal cover.” What do you mean by alignment?
Line 770-73. Mention of the SINAC PRONAMEC study keeps popping up in the discussion in the middle of discussion of biological results even though it is focused on management. I suggest you have a separate section focused on management and you put all of that discussion in there to keep the biological sections and management sections more focused. You can just basically hold the management observations until the last section of the discussion before the Conclusions.
Line 778. Change to “inhibited”.
Line 779-82. The discussion of bleaching and thermal refuge here seems redundant with the earlier discussion.
Line 783. “further stressed the coral…”
Line 786. Delete “provided”
Line 792-795. Alternative stable states again. I don’t think your data suggest anything beyond a regime shift as you have no data to suggest that the system has hysteresis. You could say that it may that there are positive feedbacks that will limit recovery, but avoid saying anything about the ecological structure of the system and its feedbacks overall since you did not test this with the observational data.
Species-specific patterns. Since Pocillopora seemed to do the worst this might be a good opportunity to contrast this with other studies in the ETP showing the some strains of Pocillopora are becoming resistant to current levels of ocean warming. Is this because Isla Caño was spared earlier bleaching and sensitive genotypes were not culled and/or endosymbionts have not been shuffled? Seems like an interesting discussion point.
Palacio-Castro AM, Smith TB, Brandtneris V, Snyder GA, van Hooidonk R, Maté JL, Manzello D, Glynn PW, Fong P, Baker AC (2023) Increased dominance of heat-tolerant symbionts creates resilient coral reefs in near-term ocean warming. ProcNatlAcadSciUSA 120:e2202388120
Romero-Torres M, Acosta A, Palacio-Castro AM, Treml EA, Zapata FA, Paz-García DA, Porter JW (2020) Coral reef resilience to thermal stress in the Eastern Tropical Pacific. Global Change Biology 26:3880-3890
Line 831. Guzman & Cortés, 1992. Also update in the References.
Lines 890-898. The immune system function paragraph seems to get pretty detailed, but isn’t really related specifically to the study. Can’t this just be wrapped up in the section before on potential strategies? Especially since you spend the last part of the preceding paragraph explaining that expensive restoration strategies are not that feasible in Costa Rica.
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.