Spatial Prioritization for the Zonation of a Reef System in a New Remote Marine Protected Area in the Southern Gulf of Mexico

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript addresses a highly relevant and timely issue: integrating remote sensing, seascape metrics, and spatial prioritization tools for managing a newly established marine protected area. The topic is of clear significance for marine spatial planning, and the datasets employed (PlanetScope imagery, in situ surveys, classification outputs) have strong potential to inform conservation decisions.
However, in its present form the manuscript requires major revisions before it can be considered scientifically sound and ready for publication. Several issues limit reproducibility, transparency, and overall clarity:
Methodological inconsistencies. The number of field sites is reported differently in various sections (45 in the Abstract, 44 in Methods, 35 in classification). This must be reconciled with a table listing all sites, their roles (training vs. validation), and associated metadata.
Incomplete methodological detail. Critical components such as the exact β-diversity coefficient, the statistical tests applied to DeltaCue change detection, and the rationale for the "127 we "ght com "inatio "s" in th" weight" d over "ay analysis are insufficiently described. The analysis cannot be reproduced without precise formulas, definitions, and justifications.
Accuracy assessment. Only overall accuracy and κ are reported. Class-specific users and producers should be provided the ability to meet accepted standards, ideally accompanied by a confusion matrix in the Supplementary Materials.
Figures and tables. Some figure captions contain minor errors. For example, in Figure three, the unit '1unit' m'unitkm'unit'or'ectcor'ected. 09' m². Tab'es Tab'ese expanded or supplemented to include classification rules (for habitat suitability) and detailed accuracy results.
Uncertainty and limitations. While uncertainty is addressed in the Supplementary Materials, it should be discussed more explicitly in the main text. An apparent limitations section is also needed, covering depth constraints, class aggregation, temporal mismatch of imagery, and potential bias in habitat detection.
Benchmarking and justification of methods. Given the known limitations of weighted overlay analysis, the study would be considerably strengthened by either benchmarking the results against alternative tools such as Marxan or Zonation or providing a robust justification for why WOA was preferred.
Supplementary data availability. Some supplementary tables and figures are either missing or incomplete. Data and scripts are currently only "available" on request; to meet open science standards, they should be deposited in a public repository with a DOI.
The study has potential, but these methodological, presentational, and clarity issues must be addressed before the paper can make a reliable scientific contribution. Improving these aspects will substantially increase the credibility, transparency, and usefulness of the work for researchers and managers.

 

Comments for author File: Comments.pdf

Comments on the Quality of English Language

The manuscript contains repeated typographical errors and awkward phrasing (e.g., “themathic,” “saptial”). A thorough language edit is needed to improve readability and professionalism.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

1. The abstract is poorly drafted. You need to tell the reader what your main research question is? How are you proposing to solve the issues? What outcome have you achieved?
2. At the end of Section 1, there is a need to provide a brief introduction concerning the overall structure of the whole paper.
3. Of the 45 surveyed sites, 35 were used for training (80%) and 10 for validation. However, the manuscript does not specify how many samples were allocated to each of the 6 habitat classes (e.g., whether rare but ecologically critical habitats like “moderate hard corals” (Class 1) had enough samples to ensure robust classification). This raises concerns about potential sampling bias—common in coral reef studies where structurally complex habitats are underrepresented—which could inflate accuracy for dominant classes (e.g., sandy beds, 45.20% of the seascape) while compromising accuracy for rare, high-priority habitats.
4. The validation relies on “training sites not used for classification” rather than a fully independent dataset (e.g., field sites not included in the initial 45 surveys). This circular validation risks overestimating classification performance, as the model may have already learned subtle patterns from the broader sampling frame.
5. The authors reduce the β-diversity calculation window from 0.2 km² (used in prior studies [8,69]) to 0.1 km² because Cayo Arenas is “relatively small” (~10 km²). However, no sensitivity analysis is provided to demonstrate how this change impacts β-diversity patterns—e.g., whether smaller windows artificially inflate spatial variability or miss broader habitat turnover gradients. Without this, the β-diversity layer (a core input to prioritization) cannot be deemed ecologically meaningful.
6. The manuscript states that “thresholds based on the data distribution” were used to identify significant benthic change (2017–2023), but provides no details on howthese thresholds were calculated (e.g., mean ± 1.96 SD, interquartile range, or machine learning-derived cutoffs). This ambiguity makes it impossible for other researchers to replicate the change detection results, a critical requirement for spatial conservation studies.
7. The authors use the meanof 127 WOA outputs to define priority areas but do not analyze: (1) which metrics drove variability in results (e.g., whether connectivity or β-diversity had the greatest influence on priority scores); (2) how often specific regions were identified as high-priority across iterations (i.e., “consensus” areas); or (3) whether extreme weight combinations (e.g., weighting benthic change most heavily) produced biologically implausible outcomes. The standard deviation map (Supplementary Material) is mentioned but not interpreted in depth, leaving readers unaware of whether high-priority areas are robust to parameter variation.
8. The reclassification of metrics to a 1–5 scale (Table 1) lacks ecological justification. For example, “no change” in benthic cover is assigned the highest score (5), but the manuscript does not explain why stable habitats are inherently more valuable than habitats with positivechange (e.g., recovery of coral cover). This assumption biases the prioritization toward static systems, contradicting the dynamic nature of coral reefs (e.g., natural recovery after disturbance).
9. The authors explicitly note that Acropora spp. and Orbicella spp.—critically endangered coral genera [116]—were observed in field surveys and drone imagery but were not included in the iterative WOA. Instead, these taxa are treated as an afterthought (“overlaying these zones with the current prioritisation map can enhance conservation planning”). This is a major oversight: the study’s priority areas may exclude microhabitats essential for the persistence of threatened corals, rendering the framework irrelevant to Mexico’s obligations under biodiversity conservation agreements (e.g., IUCN Red List targets).
10. The “habitat availability for commercial species” layer focuses on Serranidae (groupers) but does not quantify: (1) how much of the high-priority area overlaps with active fishing grounds; (2) the potential economic impact of no-take zones on local fleets; or (3) how to balance grouper habitat protection with sustainable fishing. Without this, the framework cannot guide on-the-ground management—MPAs fail if they are rejected by communities that depend on marine resources.
11. The SGMRNP (41,100 km²) is described as having “limited surveillance, funding constraints, and fragmented institutions,” but the manuscript does not explain how its 8 priority areas could be managed within these limitations. For example, are the priority areas accessible for monitoring? Can they be enforced given the MPA’s remote location? These practical questions are unanswered.
12. The statement “Data will be made available on request” is insufficient. Key spatial datasets (e.g., PlanetScope-derived habitat maps, WOA output layers, priority area shapefiles) should be deposited in a public repository (e.g., Figshare, Dryad) with persistent DOIs. Without this, other researchers cannot verify the maps or adapt the framework to other SGMRNP core zones.
13. Critical steps are underdescribed, including: (1) how bathymetry data were integrated with field depth measurements to define the <25 m study area; (2) the criteria used to select 6 frames per video for benthic cover estimation; and (3) the scripts used to calculate β-diversity in MATLAB (no code repository or syntax provided). These omissions prevent reproducibility—a foundational requirement for peer-reviewed research.
14. The framework is claimed to be “replicable for the sub-zonation within the SGMRNP,” but the manuscript does not address whether it would work for other SGMRNP reefs with different characteristics (e.g., deeper reefs, higher anthropogenic pressure). This overgeneralization undermines the study’s contribution to regional conservation.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The revised version has successfully addressed the main concerns raised in the first review. The introduction now provides the necessary background and policy context, and the methods are described with sufficient detail and transparency, including references to supplementary materials and data repositories. Figures and tables are clear and consistent, and the results are well presented and logically support the conclusions. The English has been substantially improved, and the manuscript reads fluently.

Overall, I am satisfied with the revisions, and I recommend the manuscript for publication in its current form.

Reviewer 2 Report

Comments and Suggestions for Authors

All my concerns have been fully addressed.