Remote Sensing Reveals Multidecadal Trends in Coral Cover at Heron Reef, Australia

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Review of "Remote Sensing Reveals Multidecadal Trends in Coral Cover at Heron Reef, Australia" By Carrasco-Rivero et al.

The manuscript describes in a hasty and a bit careless fashion, the creation of a time-series mapping of reef bottom types of Heron Reef, using several historical/ancillary and modern field and remote sensing imagery datasets, and the bottom types changes that are observed from the classified images.

The effort is valuable and reflects the commitment of the authors to keep alive the monitoring program in Heron Island. There are very few sites around the world that have this kind of multi-decadal datasets.

But i have some questions and a list of observations and suggestions.

Q1. In your section 2.2.3 Physical attributes, what is the reasoning behind including this information if it doesn't gets addressed again in the results or analysis?

Q2. The accuracy assessment is well implemented, but the resulting overall accuracy values of the classified images are mostly on the lower side, and the classification maps are somewhat discordant from year to year.  Would it be possible to address these caveats in the discussion section? 

Q3. The trend analysis, is a surface area and trajectories one, and in general it is ok as a preliminary analysis, but wouldn't it be more adequate and rigorous to perform a pixel per pixel analysis of change? I believe the surface trend analysis shown here, is representing both the ecological change AND the classification errors together. Also, authors do not include the analysis or discussion of the stressors mentioned in lines 113-114, regarding the trends observed.

As for the observations and suggestions:

There are 2 "Figure 1"

Section 2.2.1. So authors used data from field quadrats with at least 90% of benthic cover of the 4 selected categories. But it is not clear if the 90% is for the sum by combining these four categories, or if it is 90% dominance of a given category within each photo.

This definition is crucial to better understand the next paragraph where the authors define the assignment of photos to clases (habitats?). 

Section 2.2 /2.3 It is not mentioned if the authors carried out rigorous co-registration of images, as well as spatial resolution transformations so data could be spatially comparable.

Section 2.2.2 

Why is there a table A3 (I suppose Annex) in the middle of the mansucript... also tables must have headers, not footers.

line 166. is it missing some text at the end? 

section 2.4.2.
Line 224... That is the response variable (RV), but what are the predictor variables (PVs)?

Section 3.1 Overall Accuracy a of the classified images is very low; 76% of the classified images have and OA < 70% and this is for only 6 broad categories. Are these OA values adequate for this kind of temporal analysis??

Section 3.2.1 I wonder if the habitat classes selected are more related to the reef zonation than to the benthic cover. Especially for the coral class. 

Figure numbering jumps from figure 3 to figure 6. In figure "6" the figure caption could be redacted more clearly. And the visual comparison of trends from field vs. satellite speak volumes of the discrepancies between both.

Lines 300-308. How did Heron reef fared the massive bleaching events from 2016 and 2017 in the GBR? Did authors find a particular signature of those events in the habitats changes?

Lines 309-316 I am by no means an expert on the GBR, so this is a genuine doubt. Is it really possible that coral cover (I am assuming you mean coral cover as the area ocupied by coral habitat) can drop 54% and then bounce up 51% in two decades? Given the global degradation trends on coral reefs and given the effects of the massive bleaching events?? Also, please use correct terminology, coral cover is used as the percentage of live hard coral present in a given area. You are using the term as the area occupied by the "coral" habitat.

Section 3.2.3

Figure 7. What is the meaning of using squared values? This figure should have a panel with the GAM model fit, with confidence levels, also where is the description of the actual additive model?. 

It is not clear by the figure caption, or the methods and other results description, if it is showing the "GAM predictions derived from the use of field (RV) AND satellite data (PV)" or if it is the "GAM predictions from field data VS. the classification products data." I am still very confused about the use of the term coral cover.

Are authors saying that field data is the % amount of living hard coral in a given area, estimated through the photoquadrats anlysis, and they are plotting these values (fitting a GAM model) against the area covered by the "coral" habitat obtained from the satellite images classification? If this assumption is correct, the authors are not comparing "apples to apples" and this analysis needs to be re-considered. If my assumption is utter non-sense, authors definitely need to enhance the clarity of both methods and results sections.

Lines 327

Section 4.1

lines 360-365 may need revision depending on the data used as "coral cover"

lines 365-367 exactly which results? the ones that claim a rebound of 50% coral cover in a couple of decades? 

Section 4.2 If the focus of this section is to emphasize the ecological relevance of these maps, the bleaching events from 2016, 2017 and 2021 should be discussed here, contrasting the effects observed in the field vs. the variations in the classification of habitats, not just mentioning theoretical applications.

Conclusions
lines 491-498 But did the sampling design of the monitoring data used for these maps was adequate and focused towards producing habitat maps? Or did the authors just used what they had on hand?

 

Author Response

We thank the reviewer for taking the time to review the document and for providing such detailed and valuable feedback. We have addressed each comment, and the manuscript has been adjusted. The manuscript has improved greatly thanks to the feedback. Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript presents a valuable study combining remote sensing and field data to analyze coral cover trends over several decades. While the research topic and methods are sound, there are significant issues with the presentation and organization of the manuscript that need to be addressed before it can be considered for publication.

The authors mention that photoquadrats with less than 90% cover of relevant benthic types were excluded. However, the rationale behind this threshold is not adequately explained. The authors need to clarify how this threshold affects the representativeness of the data and whether it introduces any bias into the classification process.

The study's classification accuracy is inconsistent and needs to be addressed. The Rock/Coral class shows an alarmingly low producer's accuracy of 38%, suggesting significant issues with spectral confusion between rock, dead coral, and other benthic types. This needs to be addressed by either refining the classification methodology or providing a more thorough explanation of why this low accuracy is acceptable.

While the manuscript acknowledges the heterogeneity of coral reefs, it fails to adequately address how the physical variation in reef zones affects the accuracy of the classification. A more detailed discussion is required to explain how these differences influence data quality and classification performance.

Besides, there are numerous formatting and structural problems throughout the manuscript. For example, the figures and tables are not properly referenced or discussed in the text, and some formatting inconsistencies in the tables make them difficult to understand.

The manuscript includes two Figure 1s and two Figure 2s, which is completely confusing. These should be renamed and re-ordered to avoid this redundancy. Proper figure numbering should be maintained for clarity.

The first table in Page 6  has no headers, making it unreadable and unprofessional. Additionally, it duplicates information already presented in Figure A2, leading to redundancy. 

Table 1 in Page 10 is mentioned in the text, but it is not referenced (maybe table3?) in the manuscript. 

Where is the Tabled 3 referenced in Line258.

What is "initial f38%ield data" in Line257

Author Response

We thank the reviewer for taking the time to review the document and for providing such detailed and valuable feedback. We have addressed each comment, and the manuscript has been adjusted. The manuscript has improved greatly thanks to the feedback. Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Referee to the ms:

Remote Sensing Reveals Multidecadal Trends in Coral Cover at 2 Heron Reef, Australia

By D.E. Carrasco Rivera et al.

 

The ms  is well written and reports  the comparison of a huge amount of data collected in 20 years of field work on the Western part of Heron Island with high resolution images from satellite.

However, the conclusion of all the work doesn’t differ from the introduction where it is outlined like satellite images cannot give a detailed maps of different coral communities. The cited 21 habitat maps were reduced to 6 habitat classes very general, with mixed classes like rock/coral and sand/coral that make the proposed trends over time doubtful.

Again, the classification approach is not compared the to the classical habitat classes nomenclature (Forereef, Fringing reef, Reef flat etc.) used normally in the coral mapping (why ?), nor there is a tentative to extend the approach the remaining part of Heron island (not investigated ? no satellite images available ?)

I think that the huge amount of data collected by divers is the right approach to survey a natural area with accuracy, focusing on sensitive species and to the effects of climate change as coral mortality and changes in communities. On the contrary, the  satellite images are useful to see changes in greater areas at habitat levels.

In my opinion in the ms the great amount of field data are reduced to too few classes, wasting the possibility to do a detailed work of classification on the  high resolution satellite images. This could be a good basis for use the satellite images to survey other part of the island or other similar zones of the GBR.

 

  

 

Comments for author File: Comments.pdf

Author Response

We thank the reviewer for taking the time to review the document and for providing such detailed and valuable feedback. We have addressed each comment, and the manuscript has been adjusted. The manuscript has improved greatly thanks to the feedback. Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Thanks to the authors for a thorough revision of the manuscript. I found it much clearer.

I believe all the issues detected in the review were addresed adequately, both in the manuscript, as well as in the response letter.

I believe it is now ready for publication

Reviewer 2 Report

Comments and Suggestions for Authors

All issues have been adequately addressed, and the manuscript is now acceptable for publication.