Review Reports

Round 1

Reviewer 1 Report

The manuscript deals with an important topic of remote sensing in marine habitat mapping for biodiversity conservation purposes. It presents an approach and results of analysis of freely available satellite images carried out to generate a digital Atlas on the Coastal Marine Habitat Mapping of Tonga and assessment of an Ecological Quality Index (EQI) for those habitats proposed as a simple and direct indicator of value of marine environment. The latter should in turn be useful as a baseline for further analysis on the environment health status (outside the scope of this paper).
The distribution of different classes of coral reefs, seagrasses and mangroves, which are all valuable habitats with high biodiversity in this partly uninhabited and presumably poorly studied region of Tongan Archipelago is of high interest per se. But there are few concerns and suggestions I would encourage the authors to properly address before recommending paper for publication.  

MAJOR COMMENTS
1.    Authors identify priority habitats for conservation based on adapted zonation scheme and defined habitat classes, later each class (habitat type) receives a single EQI values assigned to it based on its presumed biological/ ecological value and importance for recovery. What is missing is how exactly this value itself can later be used for assessment of environment health status. It would be good to better clarify and justify the ecological and biological merits of proposed EQI. I was actually very surprised that out of many existing EQIs there was no “Ecological Quality Index” yet. But there is an “ecofunctional quality index (EQI)” for instance, developed with quite related objectives (see Fano et al. 2019 doi.org/10.1016/S0272-7714(02)00289-5).
2.    Please provide more detailed and specific description of applied methods and selected criteria used to delineate between habitat classes based solely on satellite images. The step to move from raster stack to the shape layer of coastal habitat types in focus remained particularly unclear to me, please highlight it with more instructive details to assure repeatability of your approach.
3.    It would be very interesting to see the discussion about presented piece of work in the context of another recent atlas dealing with marine biological recourses of Tonga Gassner P., Westerveld L., Fonua E., Takau L., Matoto A. L., Kula T., Macmillan-Lawler M., Davey K., Baker E., Clark M., Kaitu’u J., Wendt H., Fernandes L. (2019) Marine Atlas. Maximizing Benefits for Tonga. MACBIO (GIZ/IUCN/SPREP): Suva, Fiji. 84 pp.
4.    Transparency and repeatability of methods is missing, absolute lack of ground truthing and validation is taken too easily without outlining alternatives and justifying the acceptability of such lack, and that most of the discussion does not interpret or draw conclusions from direct results. There is no discussion on how exactly can those layers be used later for monitoring.
5.    Incorporating in the introduction and the discussion the international experience on habitat mapping e.g. from the https://emodnet.ec.europa.eu/en/seabed-habitats could increase readers understanding of the motivation and objectives of the work done here.  

MINOR suggestions and comments
L3 consider to edit title to something like “Baseline assessment of Ecological Quality Index (EQI) of the Marine Coastal Habitats of Tonga Archipelago: application for management of remote regions in the Pacific”
L23 and L42 Please define the scale more specifically here or in the introduction – e.g. large spatial scale (over 100s of km) - I would rather call it regional scale (regional-scale approach) otherwise
L77 “dispose at free repositories”?
L103 please specify recovery after what kind of disturbance is meant – recovery of communities can differ dependent of the type and magnitude of disturbance
L117 To start a monitoring program it is ..
L125-130 Please provide more details on methodology to be more understandable as stand-along paper to reduce the need for the reader to go into description of the mentioned software, but better understand how and based on what variables and criteria those supervised and unsupervised classification methods work. Were and adjusted or default parameters essential to be defined to obtain your results?
L130-134 this jump to the final cartography seems too abrupt, please provide more details on the workflow and explanation how spatial delineation was achieved, and how translation from classification of bands to habitat classes was carried out. I miss (perhaps more in the result section, or in supplementary) the particular classification results, perhaps threshold values for each band per habitat class, and any estimates of precision/accuracy.
L131 is it about one single resulting shapefile? I do not quite understand what is actually the reason to call one layer an Atlas? (even the estimate of EQI, if I understand it right, could be just a separate attribute of the same single polygon shapefile). Is that resulting shapefile available for scientific community outside Tonga?
L143-144 perhaps this information can be better referenced, some reports and or literature latter cited for each habitat class, or at least some underwater images and photos could serve the purpose of some kind of biological validation and illustration (e.g. at least one image per relevant habitat class in supplementary), in order to better evidence their biological value. Self-citing the Micronesia paper doi.org/10.3390/jmse7090316 is helpful, but does not seem fully sufficient.
L147 priority habitats
L154 potential
L156-158 consider to rephrase: e.g. In the classification procedure all reefs isolated from barrier reefs or islands from depth greater than 30 m were considered as RB. Consider to give a short explanation on how is reef crest, as part of the RB, is different from island in terms of marine habitat provision, i.e. why it is important to separate them in this study.
L161 is there a minimum (and maximum) size criteria for PR?
L174 is most sensitive
L177, L187 “between” instead of “among”
L185 perhaps “between the Subzone 1 and Subzone 3”?
L192 what are blue holes?
L225 do you mean that scores were averaged and after rounded up to get the final EQI score for each class (otherwise simply as a mean it would be with decimals in tab. 3)?
Figure 1 please make the background patterns for coral platform, sand and mud more sharp or coloured to be better distinguishable.
L257-259 English: “were uploaded”. This actually reads as if it was uploaded to some free repository for public access, which would be great and the best thing to do for the open science and the Decade of Ocean Science. If you mean just opening a shapefile in QGIS at your desktop, I would suggest to reformulate. As mentioned above, please provide more details on how results of habitat mapping were actually obtained and what those exactly are. What is its resolution, confidence etc…
L287 please correct English. What is each of 25 Tables? Or do you mean map sheets? What total area is covered by habitat mapping, what is area of each sheet, are there other layers and chapters in the Atlas, and how can the Atlas be accessed?
L323 “Eua, where only the Fringing Reef (FRI) was found”. I did not get it: isn’t in Eua there is also FR, RF, SE L, i.e. other classes according to Fig. 4? So why only FRI?
L333 e.g. “that all marine areas of the Archipelago have highest…”
L335 that instead of like
L337 revise the sentence: e.g. “to the highest value of EQI assigned to 71,42% of the total area covered by habitat mapping in this study”. What is actually the 100% here, but also in Figures 4 and 5 (please specify in the figures captions – percentage of what)? Is that the total area overall? Or per region? What are possible factors responsible for the distribution?
L347 that instead of like. Please specify how exactly can mortality events be detected? Is there an automated process, algorithm or pipeline that is or can be routinely and operationally used, is repeated assessment using similar approach is planned regularly? More words on recovery potential details with references or examples would be good.
L349 “images per year, meeting needs for several essential…”
L350 based on most important (unclear what first refers to)
L356 factor in tropical
L358 particularly useful
L360 bottoms with species-rich
L362 in remains unclear how is state seen and changes from one class to another can happen and be detected/captured/quantified to usefully inform management. At least some verbal explanation or working examples would be great.
L365 I think also in marine realm it is more clear to call it ground truthing; surveys of
L367 consideration of operational depth limitation of satellite images
L368 areas, e.g. offshore… methodologies, such as surveys by multibeam or sidescan sonar, or underwater imagery to obtain observations of biodiversity.
L370 ; coral species inhabiting…
L376 check references - [19] is Mitchell et al., there is no Lyons et al. in the reference list
L377 dispose a mapping…
L380 QGIS is just one of possible GIS software, sentence looks naive
L382 vs L385 how exactly is Atlas accessible by everyone (or is it only for restricted use)? How can I as a reader access it? Would be great to have a link here, or even better a reference to the shape file provided as supplementary material – that would greatly increase the value of the paper!
L383 in the field campaigns
It is important to highlight that real observational biodiversity data is essential to constitute a basis for ecosystem health status assessments.
Data Availability Statement is missing as not applicable, but it seems very important to clarify if results or habitat mapping are somehow available (perhaps at least on request?)  

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

This paper presents the use of maps (obtained from satellite imagery) as a conservation tool for shallow-water habitats in the Kingdom of Tonga, West Pacific. The paper is straightforward and clear. The purpose of the paper is very practical. I have some suggestions for improvement.

Line 34: island -> islands

Lines 35-36. Keywords should not overlap with title words. Delete: Ecological Quality Index,Tonga

Line 49. which contributed -> which contributed

Line 51. In tropical environments -> In tropical marine environments

Figure 1. It might be helpful to readers if schematic maps are added of how these reef types look like from above. I am puzzled by the reef bank as a structure outside the barrier reef, since the outer part is usually the barrier reef itself.

Line 256. Two commas: delete one.

Figure 2. These maps are too small to read. It is better to put one (a) on top of the other (b) and use the entire page width.

Figure 3. This map is too small to read. It is better to use the entire page width.

Line 362. However, it should be remembered -> One should keep in mind

 

Line 364-365. How is the need for correcting tidal fluctuations. How much is the tidal difference, anyway?

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Authors have carefully addressed all my comments and considerably improved the clarity of the manuscript. Thus, I believe it can and even should be published in the Remote Sensing journal.

Author Response

Dear Reviewer,

thanks a lot for you comments. We’re very happy about it.

Kind regards,

Dr Peirano