UAV-Hyperspectral Imaging to Estimate Species Distribution in Salt Marshes: A Case Study in the Cadiz Bay (SW Spain)
Round 1
Reviewer 1 Report
This revised manuscript, " UAV-Hyperspectral Imaging to Estimate Species Distribution in Salt Marshes: a Case Study in the Cadiz Bay (SW Spain)" (remotesensing-2213170), presents the application of high-resolution hyperspectral imaging using a UAV to determine the distribution of plant species in the salt marshes of Cadiz Bay. The methods used identified species-specific spectral signatures, allowing for accurate plant differentiation. The results showed the presence of Sarcocornia spp. and Sporobolus maritimus, with Sarcocornia occupying the majority of the area. In addition, use of UAV-HS techniques offers the advantage of easily customizable and monitorable salt marsh vegetation.
The introduction, materials and methods, results, and discussion sections are informative and well-written, but the manuscript is lengthy. It is recommended to shorten the manuscript to meet the guidelines of the journal "Instructions for Authors," making it clearer, smoother, and less dense for readers. The manuscript may have originally been a thesis and may need further formatting to meet the standards of a scientific article. Please double-check the English style for accuracy.
The figures and tables are in good condition, but it is recommended to:
Put keywords in alphabetical order;
Standardize equipment/reagents/software nomenclature by including the manufacturer, city, state, and country using the three-letter code;
Review the "104" references before 2010.
Best regards,
Author Response
Please see the attachment
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Currently, one of the trends in the development of remote sensing methods is the increasing use of drones (UAVs). It should be noted that the use of UAVs is in some way similar to the use of small aircraft, the method which was popular in the 1970-1990s. And it should be noted that the UAV is an incomparably cheaper than small aircraft and therefore more profitable.
It should also be noted that the topic of remote sensing of vegetation using multispectral and hyperspectral methods has been actively developing over the past 10-15 years, and remains popular at the present time. A significant amount of knowledge on this topic has been accumulated.
The main object of study is solonchaks. Currently, there are more than a billion hectares of saline soils on the globe, including solonchaks. There is an opinion that the biological productivity of solonchaks and highly saline soils is very low. Accordingly, these lands are rarely used for agriculture. At the same time, there is a need to develop methods for remote sensing of such natural objects.
Based on the foregoing, we can conclude that the topic of the article is relevant and timely.
The article itself is interesting, scientifically based, written in an understandable scientific language, well structured, illustrated, methodically verified.
Note that figure 2 is repeated 2 times (page 4 and 5).
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
The article is of interest for the development of methods for monitoring plant communities. The article is methodically verified, the conclusions of the authors are sufficiently substantiated. At the same time, there are a number of comments and questions to the article.
The text is very difficult to understand, overloaded with terminology and will be perceived only by narrow specialists. It can be proposed to shorten the descriptions of the technical characteristics of the devices, since they are publicly available.
Hyperspectral imaging was carried out once without repetition. This may cast doubt on the results obtained.
The authors did not conduct independent field studies, but used the measurements of the spectra of the FAST project fields as a verification data, using them as a reference library of spectra to verify the spectral characteristics obtained using the UAV-HS classification (2014–2015). What guarantees the combination of data from two different studies?
It is known that random forest often shows better results than SVM. How much better is the presented method than RF?
Specific comments:
Lines 24. “Two varieties of Sarcocornia spp.” should be changed to “Two species of Sarcocornia spp.”.
Lines 25, 82. It is necessary to indicate the author of the botanical taxon Sporobolus maritimus.
Line 37. Keywords should not contain terms from the title of the article. Remove «salt marshes» from the keywords and add a term that is not present in the title.
Line 88. It is necessary to indicate the author of the botanical taxon Spartina anglica.
Line 89. It is necessary to indicate the author of the botanical taxon Sporobolus anglicus.
Line 92. “UAVs that fly up to 120 m altitude are suitable to identify spatial heterogeneity in microtopography canopy height or greenness.” Are you planning to use a hyperspectral or RGB camera for these purposes?
Line 100. It is necessary to indicate the author of the botanical taxon Spartina alterniflora.
Line 155. The authors of the botanical taxa Sporobolus fruticose, Sporobolus perennis, Zostera noltei, Cymodocea nodosa and Zostera marina should be given.
Line 265. Why was the 0.3 threshold chosen?
Line 475. A very low classification accuracy (46%) was obtained.
Line 500. “Error! Reference source not found”. What is this?
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Dear author, I recommend the manuscript. Best regards
