GIS-Based Assessment of Habitat Networks for Conservation Planning in Kas-Kekova Protected Area (Turkey)

Round 1

Reviewer 1 Report

When asked to review this work for a second time, I was optimistic that the authors had addressed the serious flaws with the first submission and that the manuscript had been significantly improved. I am appreciative of the time and energy that the authors have put into revision, and the question addressed by this manuscript is still interesting and important. However, several of the key points raised in my previous review (as well as Reviewer #3 from the previous submission) were not addressed in a satisfactory fashion. Because these fundamental flaws remain, I regret that I must recommend this manuscript be rejected by the journal. At this point, I admit I am skeptical that the authors will be able to resolve these fundamental issues with the manuscript without a substantial amount of additional work.

 

Specific comments below:

 

Although the authors have apparently sent the manuscript to a professional English language service, several grammatical issues remain and still impair the readability of this manuscript. These must be corrected prior to publication. For instance, “Although various spatial analysis methods can be utilized for habitat pattern determination purposes, the method selection and parameterization should be performed carefully especially since the threatened species can be observed rarely and using only few observations may not reveal the whole pattern” (Lines 187-189). This statement is not coherent. Key aspects of the sampling methodology are inadequately described. The changes made to address this point are unsatisfactory. For instance, “The extensive field works held by specialists throughout the project (four consecutive spring and fall seasons in two years) ensured the representative data collection for biodiversity” (lines 121-123). This is not an adequate description of spatial sampling methodology. The justification of the kernel width choice is unsatisfactory. The response of “We have actually tested other bandwidths (500 m, 800m, 1000m) and found 1 km visually suitable” is not adequate justification. The experiment of varying kernel width included in the review response is interesting and would add considerably to the scientific content of the manuscript without lengthening it unreasonably. The answer to my previous question about topographic barriers is unsatisfactory. The authors state in their review response “We have analyzed the topography in terms of altitude ranges and slope categories added to the manuscript (Figures 7,8,9). They do not indicate any topographic barriers to the mammals and reptiles and although the aspect value is important for threatened flora, we did not need to analyze it any further.” The text associated with Figures 7, 8 and 9 (“no clear indications of spatial correlation with the topographical characteristics were observed, therefore, the topography was not found as an effective barrier for the biodiversity elements”) is inadequate.  

Author Response

Dear Reviewer,

 

Thank you very much for reviewing our revised manuscript. We highly appreciate your contributions.

 

We have sent the manuscript to the English editing service of MDPI before this revision. We hope that the new revisions we have made may satisfy you. Please see our replies to your comments in the attached file.

 

Kind regards,

Authors

Author Response File: Author Response.docx

Reviewer 2 Report

A final touch on the text is needed. The article shouldn't promise but it should provide the clear messages oriented towards researchers, conservationists, biodiversity monitoring experts, planners, regional decision makers, national governments, etc.

Proofreading by a native English speaker is needed as there are still small but numerous grammatical and spelling errors left and occuring anew.

Some non-exhaustive concrete comments can be proposed at:

Line 14 > (biodiversity data) and assessment Line 25 > (for land use planning) and biodiversity monitoring (in the PA) Line 106 > The study area of 260 km2 includes both aquatory and territory. The area of the latter is available in Table 6 but should be clear from the beginning Figure 2 > The administrative districts are several seperate settlements' areas which are part of two districts Kaş and Kale, how many municipalities? The corresponding human population is from which reference year? Line 428 > maybe prohibition is not the only approach as agro-ecological measures can be an option in some cases Line 432 > (construction) of (new roads) and reconstruction of exhisting Line 444 > "expert opinion is required for kernel radius selection" maybe should be changed to more expertize and field studies are needed for aquisition of more knowledge about the threatened species.

 

 

Author Response

Dear Reviewer,

 

Thank you very much for reviewing our revised manuscript. We highly appreciate your contributions.

 

We have sent the manuscript to the English editing service of MDPI before this revision. We hope that the new revisions we have made may satisfy you. Please see our replies to your comments in the attached file.

 

Kind regards,

Authors

Author Response File: Author Response.docx

Reviewer 3 Report

This paper has been fully revised and can be accepted for publication.

Author Response

Dear Reviewer,

 

Thank you very much for reviewing our manuscript once again. We highly appreciate your contributions to our work.

 

Kind regards,

Authors

Round 2

Reviewer 1 Report

To my surprise, the authors have satisfied the most important of my concerns. The scientific quality of the manuscript is greatly improved with the addition of the kernel density test and the more explicit and coherent justification of the decisions made during the analysis. I am happy to recommend that the manuscript be published.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

This manuscript attempts to assess habitat networks in the Kas-Kekova Protetcted Area of Turkey for the purposes of conservation planning. The study is framed in the context of Aichi target 11, which defines a goal for the global establishment of protected areas. As I understand it, this manuscript takes existing datasets and applies kernel density smoothing, followed by a Jenks classification to map areas on the basis of habitat criticality for endemic flora, reptiles/amphibians, mammals, and birds. The maximum Jenks category from each of these biota classes is then taken as the final map, which is overlaid on a map of land ownership/use to determine importance for protection.

 

Mild to moderate English language issues impair the readability of this manuscript. It would benefit from proofreading by a native English speaker. Possibly because of language issues, fundamental aspects of the analysis remain unclear to me. For instance, what was the spatial sampling approach to obtaining the points of endemic/endangered flora & fauna in Figure 7? Are some areas badly undersampled? If so, which ones? Were other kernel widths tested? What is the sensitivity of the output map to kernel width? Is there reason to expect that a uniform kernel might not give an appropriate answer (e.g. do steep topographic gradients exist that break up what would otherwise appear as uniform corridors)?

 

In my opinion, the question addressed by this manuscript is interesting and important. However, these outstanding fundamental issues are sufficient to preclude publication in current form. Specifically, the explanation and justification of the analysis approach and methods used must be clarified and expanded. It seems likely to me that the authors can address these concerns without fundamentally changing their approach. However, significant additional analysis would be required that probably could not be done on the timeline of a major revision for this journal. I suggest the manuscript be rejected and resubmitted after the authors have taken the time to more fully address these fundamental concerns.

 

 

Line-specific comments below:

 

Line 60: “For example, the distribution of plant species can be expressed with discrete points”. Would it not take thousands of points to accurately represent the spatial (and temporal) variations in plant species over any appreciable area? Don’t most efforts use polygons which attempt to map areas on the basis of dominant plant species composition?

 

Figures 1 and 2. I suggest identifying the location of Figure 2 with a point on Figure 1.

 

Figure 3. I suggest including vector outlines of the settlements in addition to the circle representing population size.

 

Figure 4. What does the red polygon show in this figure?

 

Line 157: “first the flora and fauna point data collected in the field work [36] were analyzed to determine patterns of plant and animal ecology and to produce ecological sensitivity maps [43].” The reader would greatly benefit from having some idea of where these point data are actually located in geographic space.

 

Figure 6: I suggest modifying the color scheme for this figure. For instance it is difficult to visually distinguish Maquis, Low-mid saltmarshes, and mixed Mediterranean pine.

 

Figure 7: How is this map derived? Are these actual field observations of endemic/threatened species? Or based on GIS overlay/habitat suitability? Is this new from this study or derived from previous work? What do the polygons represent? Are these just the same polygons from Figure 6?

 

Lines 218-230: On what data was KDE applied? Is this just smoothing the point data from Figure 7? How do we know these points are representative? What about sampling bias?

 

Line 225: Why was 1km used? Were other radii examined as well? What is the sensitivity to kernel radius?

 

Lines 246-250: I suggest making clearer why the “continuous habitat pattern” is biogeographically real and not just a function of the geographic location of sampling points and the width of the kernel used.

 

Lines 250-251: I suggest expanding on the point of incomplete spatial sampling and providing a more complete discussion of its implications and locations of particular concern for undersampling.

 

Figure 10: The way I understand this map, it takes the maximum Jenks value from any of the maps in Figure 9 (a-d). However, from Fig 9, it appears some areas area at the highest importance level for more than one biota class. Might it be interesting to examine these locations and list them at even higher conservation priority? Or is this incorporated in the map already?

Comments for author File: Comments.pdf

Reviewer 2 Report

The article reflects two of the themes in the scope of the journal, namely "spatial analysis ..." and "applications of geoinformation technologies".

The content is significant, especially due to the importance of the object of study - the area protecting rich Meditteranean biodiversity and the role of  geoinformation technologyies to inform sound spatial planning practice and decision making in the region.

Several comments can be made about the improvement of the research work, regarding the research design and methods used, the presentation and interpretation of the results and in relation to both of these the relevant conclusions:

Research design and methods

The analytical methods (KDE) and processing (Natural breaks) are not innovative and have often been applied in other research, experiments and practice with biodiversity and other datasets with reviewing of limitations of the KDE. The research design can be improved in both ecosystems and geoinformatics directions through more rich and integrated approach.  The use of habitat suitability or ecological niche modelling through GIS, especially for the amphibians, mammals and birds is important as these species are not static and they have more complex areal of feeding, nesting, etc. The estimation of barriers - man-made or other natural features which occur at the experimental area can also have significant effect on biodiversity corridors which has to be taken into account in the application of the method.

Presentation of the results

The presentation of the results through overlaying of property cadastral data with the core habitats and buffer patches can be improved in terms of cartographic methods (lighter gray of private property, stronger colors of the habitats (less transparency) or use of contour lines (isophlet) instead). It is not clear if the percentage of different ownership types is for the whole or only for the core habitats and buffer pathes and there is no parameter for the area of the whole PA in the introduction or in the main characteristics parts.

The part commenting the conservation strategies should make more clear what the contribution of the authors is beyond the existing strategies of the MoEU.

Conclusions

Part of the results which are arbitrary and need deeper studies, aquisition of data or stakeholder involvement through the use of geoinformation technologies can be further explained in the final chapter. This will provide more focused impression about the way these technologies can contribute to the pressing need of better adapted planning regimes for development and protection of this precious area. Future works are mentioned but they are broadly commenting other activities and are not focused on the applicaiton of geoinformation technology. Does it mean that what has been done as spatial analysis is enough for the future land use planning and management practice in this case?

Reviewer 3 Report

The authors present the creation of a map from input data using standard classification procedures.While the topic in general might be interesting to the readers I see the following issues:

The data sets are only described in a superficial way. The audience of the journal is not habitat experts, thus we cannot expect them to understand a simple list of habitat types and how they are differentiated. It is also insufficient to specify point as format and provide a spatial accuracy. How are the points distributes, is it allowed to interpolate between them (there is a difference between a DTM represented as points and a collection of trees represented as points, for example). The methods used to produce the results were theoretically explained but it remains unclear why these steps were performed. The process how to derive the habitat classes from the input data on a conceptual level (not the functional level of the GIS) is not explained at all. Why was the KDE method use, what is the reason for the radius of 1km, why was the Jenks natural break classification used, why 5 categories, and what is the reasoning behind the combinatory approach (it eliminates all information except for the dominant one). The final map looks very colorful but does it communicate what it shall? Was this tested with users?

The paper leaves the impression that the authors had data, applied some standard methods (KDE, Jenks classification), and published the results. There is no reasoning that others can use in their own analysis.

Some small comments:

p. 2 Please avoid judgments like "is immensely required"

Base maps cannot serve as scientific evidence. It is the data represented in the maps that is the scientific evidence. But maps can lie quite easily (compare Mark Monmonier, "How to lie with maps")