Monitoring Ecological Conditions by Remote Sensing and Social Media Data—Sanya City (China) as Case Study

Round 1

Reviewer 1 Report

This study consolidates the idea that it is not enough for decision-makers to exclusively use the information data coming from satellite images! Complementary data, looking the people perception/sentiments on different phenomena and on environment quality, are necessary for a better territorial management.

The authors proposed a paper, demonstrating the importance of remote sensing and social media for a better monitoring of ecological environment. Continuing some research in the field, the authors offer explicitly methodological steps for such analysis and encourage the scholars to use multi-source data in their projects.

I highly appreciate the general structure of the paper, which allocates more space to methodological issues and optimum one to results. I underline the clarity of the used methods, and the rigor of the general analysis reflected in the obtained results and discussions.  

General comments and suggestions:

1. Please reflect if is not the case to make, sometimes, a distinction between “public sentiment information” and “public perception”! I know that you use the Sentiment Analysis Method, but in some contexts it’s about public perception! Just for reflection!!!
2.  Regarding the paper’ structure, even if I agree it, I suggest adding to the Results section title and Discussion! Why? Because, practically, you comment/discuss the obtained results, coming with useful explanations. 

Specific suggestions or corrections:

1)      Improving of the paper title, highlighting the complementarity of remote sensing and social media in the monitoring process of the ecological environment. By consequence, I propose a small change: Monitoring the ecological environment by remote sensing and social media. Sanya city (China) as case study. In my opinion, considering the importance of the methodology, and the fact that the case study serves only to demonstrate its efficiency, I would like to eliminate the case study from the title. But you decide!

2)      Figure 1 is unclear! This should be improved, because using only black and white it’s difficult for reader to have a real image on the land use structure of the city.

Some technical remarks:

- replace “researches” with “research”;

- replace “link relationship” with “link” or “relationship”;

- eliminate the second article from some formulations as: “the higher the….”, “the smaller the…”, the thicker the…” etc;

- improve the first paragraph structure from Conclusion section, which starts with “We took Sanya City as an example to verify the effectiveness of this method and found that 1).............2)”;

- in the second paragraph of Conclusion section it’s required „to” in the context of „can help decision makers (to) understand the changing.....”. 

- replace the last words of the Conclusion section „of using the data” with „of using data”.

Author Response

Thank you very much for your recognition of our work and your valuable comments. I am very confident that these opinions will make this article better.

• General comments and suggestions:
• Please reflect if is not the case reflect if is not the case to make, sometimes, a distinction between “public sentiment information” and “public perception”! I know that you use the Sentiment Analysis Method, but in some contexts it’s about public perception! Just for reflection!!!

The question you provided is very good and worth thinking about. In my opinion, the two terms are relatively similar, but there are some differences. "public sentiment information" is more specific and covers a narrower scope. It simply refers to the public's intuitive emotional tendencies towards a certain theme, concept, etc. The meaning of "public perception" is more abundant. In addition to including emotions, it also includes some ideas and opinions of the public, which may be difficult to describe intuitively using sentiment analysis.

In this paper, we used sentiment analysis to categorize social media data explicitly, and then further process the classified data to perceive the public's thoughts on the ecological environment.

• Regarding the paper’ structure, even if I agree it, I suggest adding to the Results section title and Discussion! Why? Because, practically, you comment/discuss the obtained results, coming with useful explanations.

Thank you for approving the structure of this paper. We have made changes to the title of Section 4. In this paper, we analyze the ecological environment of the study area from two aspects, namely Section 4.1 and Section 4.2. In these sections, we first describe and analyze the results calculated by the method in this paper in detail. Then, we discussed and concluded the advantages of this method in monitoring the ecological environment. Therefore, it is more reasonable to change the title of Section 4 to "Results and Discussion".

• Specific suggestions or corrections:
• Improving of the paper title, highlighting the complementarity of remote sensing and social media in the monitoring process of the ecological environment. By consequence, I propose a small change: Monitoring the ecological environment by remote sensing and social media. Sanya city (China) as case study. In my opinion, considering the importance of the methodology, and the fact that the case study serves only to demonstrate its efficiency, I would like to eliminate the case study from the title. But you decide!

Thanks for your suggestion, I think it's good. In addition, I was also inspired by your comments, how do you feel about changing the title to "Building social semantic networks to assist remote sensing for ecological monitoring—Sanya city (China) as case study ". Regarding the description of the case study in the title, this may need to be added because of the associated project funding.

• Figure 1 is unclear! This should be improved, because using only black and white it’s difficult for reader to have a real image on the land use structure of the city.

Figure 1 is really not clear enough, so we changed it.

• Some technical remarks:

Thanks for pointing out so many detailed errors in the paper. We corrected these errors individually and invited professional English editors to correct grammatical errors, thereby improving the readability of this paper.

Thank you again for your contribution to our manuscript, and your comments will also help our follow-up research.

Author Response File: Author Response.docx

Reviewer 2 Report

Reviewer Comments

Journal: remote sensing

Manuscript Title: “Monitoring ecological environment combined with remote sensing and social media -- a case study of Sanya, China” 

General Comments: The authors present a new approach to combing social and ecological data to identify locations with ecological vulnerability.  Overall all, the paper is interesting and includes several impressive graph figures. 

The overall analysis appears appropriate and unique, but additional detail is needed to fully understand all aspects of the methods.  A few comments are provided below as examples.

Also, the manuscript did not include line numbering.  Because of this, few grammatical suggestions were included in this review.  However, an English language check is suggested.

Comments:

Abstract: Suggest removing “the node” as it has not yet defined.  Stick with just “spatial location”.

First paragraph graph of Intro:

1.     “et al” or “etc” (etcetera)?  (this issue is recurrent in several places)

2.     “traditional statistics” as compared to AI or machine learning?

Section 2.2.1: Suggest listing all data and not using “etc”

Section 2.2.2: Need to list all data, do not use “etc” here.

Section 3.1.1: Please report the actual weights used for each of the indicators.

Section 3.1.2: What specific GIS software, tool, and model were used?

Section 3.1.2: What numeric thresholds or criteria were used to divide the area into different categories?

Section 3.4.1: This section provides a good conceptual overview of what was done to combine ecological and social data, but it is unclear how each of these data sets were weighted relative to each other.  Did the social and ecological each contribute 50% to the combined analysis, or was one type weighted more highly?

Section 3.4.2: How were larger nodes given “special attention”?

Fig 4: Suggest using a 2D pie chart rather than a 3D version for print.

Great job with the network diagrams!  It is often difficult to display such large networks, but these figures look good and are easy to interpret.

Author Response

Thank you for your valuable comments that make our work better. We have carefully considered these comments and revised the paper accordingly.

1. Also, the manuscript did not include line numbering.  Because of this, few grammatical suggestions were included in this review.  However, an English language check is suggested.

Thanks for pointing out the inadequacies. Regarding the language issue, we have invited professional English editors to revise the manuscript as a whole, which will be more conducive to reading.

2. Abstract: Suggest removing “the node” as it has not yet defined.  Stick with just “spatial location”.

Sorry, we did not express clearly here, we have made corresponding changes.

3. “et al” or “etc” (etcetera)?  (this issue is recurrent in several places)

We checked the relevant explanations, and in general, “et al” is mainly used to omit the person mentioned, “etc” refers to things, events, etc. Therefore, we made careful modifications.

4. “traditional statistics” as compared to AI or machine learning?

I'm so sorry we didn't write clearly here. What the manuscript is trying to convey is “statistical data”.

5. Section 2.2.1: Suggest listing all data and not using “etc”

Section 2.2.2: Need to list all data, do not use “etc” here.

We have revised the corresponding question, listing all the data in detail, and removed "etc".

6. Section 3.1.1: Please report the actual weights used for each of the indicators.

We added a table in the corresponding position of the manuscript, which records the weight of each indicators in detail.

7. Section 3.1.2: What specific GIS software, tool, and model were used?

Sorry we didn't describe it clearly here. We used ArcGIS software and calculated the weighted sum of all indicators through a multi-factor weighted summation model.

8. Section 3.1.2: What numeric thresholds or criteria were used to divide the area into different categories?

We used the natural break method in ArcGIS software to divide the calculation results, and we revised the description of the original text to make the expression of the method clearer.

9. Section 3.4.1: This section provides a good conceptual overview of what was done to combine ecological and social data, but it is unclear how each of these data sets were weighted relative to each other.  Did the social and ecological each contribute 50% to the combined analysis, or was one type weighted more highly?

This question is very valuable. In this paper, we actually downplay the concept. The method that we provided is being integrated into an actual application software to serve the ecological environment monitoring in Sanya, China. In practical applications, how to determine the weight of different data may depend on the user's subjective judgment. We mainly provide a qualitative analysis. We provide relatively important areas, which are often of high ecological sensitivity and public concern. Generally speaking, social media and ecological sensitivity each contributed 50% to the final analysis. Of course, there are also some areas, based on social media data, it is found that the local ecological environment is seriously damaged, but the ecological sensitivity may not be the highest in this area. We will also focus on these areas.

In future research, we are also thinking about how to reasonably quantify different data. We are also trying to find a model that can fuse these quantitative data more efficiently. But since social media is a kind of unprofessional data generated spontaneously by the public, the work can be difficult. However, it is well worth trying. In this paper, we made a preliminary attempt and also got the results we wanted, which has great encouragement for future research.

10. Section 3.4.2: How were larger nodes given “special attention”?

We didn't describe this place clearly. What we actually want to express is that for those larger nodes, we can further analyze the reasons why the public pays more attention to the corresponding areas. We have modified this.

11. Fig 4: Suggest using a 2D pie chart rather than a 3D version for print.

We changed the original image to 2D, which is really much better.

Thank you again for your contribution to our manuscript, and your comments will also help our follow-up research.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Excellent job with the revisions.  All of my concerns have been addressed.